Pasar al contenido principal

METODOS ESTADISTICOS FLEXIBLES EN CIENCIA DE DATOS PARA DATOS COMPLEJOS Y DE GRAN VOLUMEN: TEORIA Y APLICACIONES

PID2020-113578RB-I00

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal de I+D+i Orientado a los Retos de la Sociedad
Convocatoria Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Año convocatoria 2020
Plan Estatal 2017-2020 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSIDADE DA CORUÑA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 64
Encontrada(s) 1 página(s)

e-Archivo. Repositorio Institucional de la Universidad Carlos III de Madrid
Publicación Artículo científico (JournalArticle)

Bootstrap bandwidth selection and confidence regions for double smoothed default probability estimation

  • Peláez Suárez, Rebeca
  • Cao, Ricardo
  • Vilar, Juan M.
This article belongs to the Special Issue Application of Survival Analysis in Economics, Finance and Insurance, For a fixed time, t, and a horizon time, b, the probability of default (PD) measures the probability that an obligor, that has paid his/her credit until time t, runs into arrears not later that time 𝑡+𝑏. This probability is one of the most crucial elements that influences the risk in credits. Previous works have proposed nonparametric estimators for the probability of default derived from Beran’s estimator and a doubly smoothed Beran’s estimator of the conditional survival function for censored data. They have also found asymptotic expressions for the bias and variance of the estimators, but they do not provide any practical way to choose the smoothing parameters involved. In this paper, resampling methods based on bootstrap techniques are proposed to approximate the bandwidths on which Beran and smoothed Beran’s estimators of the PD depend. Bootstrap algorithms for the calculation of confidence regions of the probability of default are also proposed. Extensive simulation studies show the good behavior of the presented algorithms. The bandwidth selector and the confidence region algorithm are applied to a German credit dataset to analyze the probability of default conditional on the credit scoring., This research has been supported by MICINN Grant PID2020-113578RB-100, and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro Singular de Investigación de Galicia ED431G 2019/01), all of them through the ERDF.
Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela
Publicación Artículo científico (JournalArticle)

Nonparametric estimation for a functional-circular regression model

  • Meilán-Vila, Andrea
  • Crujeiras Casais, Rosa María
  • Francisco Fernández, Mario
Changes on temperature patterns, on a local scale, are perceived by individuals as the most direct indicators of global warming and climate change. As a specific example, for anAtlantic climate location, spring and fall seasons should present amild transition between winter and summer, and summer and winter, respectively. By observing daily temperature curves along time, being each curve attached to a certain calendar day, a regression model for these variables (temperature curve as covariate and calendar day as response)would be useful for modeling their relation for a certain period. In addition, temperature changes could be assessed by prediction and observation comparisons in the long run. Such a model is presented and studied in this work, considering a nonparametric Nadaraya–Watson-type estimator for functional covariate and circular response. The asymptotic bias and variance of this estimator, as well as its asymptotic distribution are derived. Its finite sample performance is evaluated in a simulation study and the proposal is applied to investigate a real-data set concerning temperature curves.
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Comments on: Nonparametric estimation in mixture cure models with covariates

  • Cao, Ricardo
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature., [Abstract]: This paper discusses the invited paper by López-Cheda, Peng and Jácome on nonparametric mixture cure models with covariates. An alternative estimation procedure is proposed in this context. The situation when the two covariate vectors (the one in the incidence and in the latency parts) share some, but not all, their covariates is also considered. Some technical aspects in the assumptions, results and proofs of the invited paper are also discussed. Comments on the simulations and the real-data application are included. Finally, possible interesting topics for further research in this field are briefly discussed., This research has been supported by MICINN Grant PID2020-113578RBI00 and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro de Investigación del Sistema Universitario de Galicia ED431G 2019/01), all of them through the European Regional Development Fund (ERDF). Funding for open access charge: Universidade da Coruña/CISUG., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Modeling the Number of People Infected With SARS-COV-2 From Wastewater Viral Load in Northwest Spain

  • Vallejo, J. A.
  • Trigo Tasende, Noelia
  • Rumbo-Feal, Soraya
  • Conde-Pérez, Kelly
  • López-Oriona, Ángel
  • Barbeito, Inés
  • Vaamonde, Manuel
  • Tarrío-Saavedra, Javier
  • Reif López, Rubén
  • Ladra, Susana
  • Rodiño-Janeiro, Bruno Kotska
  • Nasser-Ali, Mohammed
  • Cid, Ángeles
  • Veiga, María Carmen
  • Acevedo, Antón
  • Lamora, Carlos
  • Bou, Germán
  • Cao, Ricardo
  • Poza, Margarita
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, [Abstract] The quantification of the SARS-CoV-2 RNA load in wastewater has emerged as a useful tool to monitor COVID–19 outbreaks in the community. This approach was implemented in the metropolitan area of A Coruña (NW Spain), where wastewater from a treatment plant was analyzed to track the epidemic dynamics in a population of 369,098 inhabitants. Viral load detected in the wastewater and the epidemiological data from A Coruña health system served as main sources for statistical models developing. Regression models described here allowed us to estimate the number of infected people (R2 = 0.9), including symptomatic and asymptomatic individuals. These models have helped to understand the real magnitude of the epidemic in a population at any given time and have been used as an effective early warning tool for predicting outbreaks in A Coruña municipality. The methodology of the present work could be used to develop a similar wastewater-based epidemiological model to track the evolution of the COVID–19 epidemic anywhere in the world where centralized water-based sanitation systems exist., This work was supported by EDAR Bens S.A., A Coruña, Spain [grant references INV04020, INV12120 and INV05921 to MP], the National Plan for Scientific Research, Development and Technological Innovation 2013-2016 funded by the ISCIII, Spain - General Subdirection of Assessment and Promotion of the Research-European Regional Development Fund (FEDER) “A way of making Europe” [grant numbers PI15/00860 to GB, PI17/01482 and PI20/00413 to MP], the GAIN, Xunta de Galicia, Spain [grant number IN607A 2016/22 to GB, ED431C-2016/015 and ED431C-2020/14 to RC, ED431C 2017/58 to SL, ED431G 2019/01 to RC and SL, and ED431C 2017/66 to MCV], MINECO, Spain [grant number MTM2017-82724-R to RC], Ministerio de Ciencia e Innovación, Spain [grant number PID2020-113578RB-100 to RC], and the Spanish Network for Research in Infectious Diseases [REIPI RD16/0016/006 to GB]. The work was also supported by the European Virus Archive Global (EVA-GLOBAL) project that has received funding from the European Union's Horizon 2020 - Research and Innovation Framework Programme under grant agreement no 871029. SR-F was financially supported by REIPI RD16/0016/006, KC-P by IN607A 2016/22 and the Spanish Association against Cancer (AECC) and JAV by IN607A 2016/22. Funding for open access charge: Universidade da Coruña/CISUG, EDAR Bens S.A.; INV04020, EDAR Bens S.A.; INV12120, EDAR Bens S.A.; INV05921, Xunta de Galicia; IN607A 2016/22, Xunta de Galicia; ED431C-2016/015, Xunta de Galicia; ED431C-2020/14, Xunta de Galicia; ED431C 2017/58, Xunta de Galicia; ED431G 2019/01, Xunta de Galicia; ED431C 2017/66
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Mapping the Poverty Proportion in Small Areas under Random Regression Coefficient Poisson Models

  • Diz-Rosales, Naomi
  • Lombardía, María José
  • Morales, Domingo
Cursos e Congresos, C-155, [Abstract] In a complex socio-economic context, policy makers need highly disaggregated poverty indicators. In this work, we develop a methodology in small area estimation to derive predictors of poverty proportions under a random regression coefficient Poisson model, introducing bootstrap estimators of mean squared errors. Maximum likelihood estimators of model parameters and random effects mode predictors are calculated using a Laplace approximation algorithm. Simulation experiments are conducted to investigate the behaviour of the fitting algorithm, the predictors and the mean squared error estimator. The new statistical methodology is applied to data from the Spanish survey of living conditions to map poverty proportions by province and sex, developing a tool to support policy decision making, Xunta de Galicia; ED431C-2020/14, This research is part of the grant PID2020-113578RB-I00, funded by MCIN/AEI/10.13039/501100011033/. It has also been supported by the Spanish grant PID2022-136878NB-I00, the Valencian grant Prometeo/2021/063, by the Xunta de Galicia (Competitive Reference ED431C-2020/14) and by CITIC that is supported by Xunta de Galicia, collaboration agreement between the Consellería de Cultura, Educación, Formación Profesional e Universidades and the Galician universities for the reinforcement of the research centres of the Sistema Universitario de Galicia (CIGUS). The first author was also sponsoredby the Spanish Grant for Predoctoral Research Trainees RD 103/2019 being this work part of grant PRE2021-100857, funded by MCIN/AEI/10.13039/501100011033/ and ESF+
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

F4: An All-Purpose Tool for Multivariate Time Series Classification

  • López-Oriona, Ángel
  • Vilar, José
This article belongs to the Special Issue Data Mining for Temporal Data Analysis, [Abstract] We propose Fast Forest of Flexible Features (F4), a novel approach for classifying multivariate time series, which is aimed to discriminate between underlying generating processes. This goal has barely been addressed in the literature. F4 consists of two steps. First, a set of features based on the quantile cross-spectral density and the maximum overlap discrete wavelet transform are extracted from each series. Second, a random forest is fed with the extracted features. An extensive simulation study shows that F4 outperforms some powerful classifiers in a wide variety of situations, including stationary and nonstationary series. The proposed method is also capable of successfully discriminating between electrocardiogram (ECG) signals of healthy subjects and those with myocardial infarction condition. Additionally, despite lacking shape-based information, F4 attains state-of-the-art results in some datasets of the University of East Anglia (UEA) multivariate time series classification archive., This research has been supported by the Ministerio de Economía y Competitividad (MINECO) grants MTM2017-82724-R and PID2020-113578RB-100, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14), and the Centro de Investigación del Sistema Universitario de Galicia, “CITIC” grant ED431G 2019/01; all of them through the European Regional Development Fund (ERDF). This work has received a discount in publication fees by Universidade da Coruña/CISUG, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Quantile-Based Fuzzy Clustering of Multivariate Time Series in the Frequency Domain

  • López-Oriona, Ángel
  • Vilar, José
  • D'Urso, Pierpaolo
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, [Abstract] A novel procedure to perform fuzzy clustering of multivariate time series generated from different dependence models is proposed. Different amounts of dissimilarity between the generating models or changes on the dynamic behaviours over time are some arguments justifying a fuzzy approach, where each series is associated to all the clusters with specific membership levels. Our procedure considers quantile-based cross-spectral features and consists of three stages: (i) each element is characterized by a vector of proper estimates of the quantile cross-spectral densities, (ii) principal component analysis is carried out to capture the main differences reducing the effects of the noise, and (iii) the squared Euclidean distance between the first retained principal components is used to perform clustering through the standard fuzzy C-means and fuzzy C-medoids algorithms. The performance of the proposed approach is evaluated in a broad simulation study where several types of generating processes are considered, including linear, nonlinear and dynamic conditional correlation models. Assessment is done in two different ways: by directly measuring the quality of the resulting fuzzy partition and by taking into account the ability of the technique to determine the overlapping nature of series located equidistant from well-defined clusters. The procedure is compared with the few alternatives suggested in the literature, substantially outperforming all of them whatever the underlying process and the evaluation scheme. Two specific applications involving air quality and financial databases illustrate the usefulness of our approach., The authors are grateful to the anonymous referees for their comments and suggestions. The research of Ángel López-Oriona and José A. Vilar has been supported by the Ministerio de Economía y Competitividad (MINECO) grants MTM2017-82724-R and PID2020-113578RB-100, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14), and the Centro de Investigación del Sistema Universitario de Galicia “CITIC” grant ED431G 2019/01; all of them through the European Regional Development Fund (ERDF). This work has received funding for open access charge by Universidade da Coruña/CISUG, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Time series clustering based on prediction accuracy of global forecasting models

  • López-Oriona, Ángel
  • Montero Manso, Pablo
  • Vilar, José
This version of the article: López-Oriona, Á., Montero-Manso, P., & Vilar, J. A. (2025). ‘Time series clustering based on prediction accuracy of global forecasting models’ has been accepted for publication in: Knowledge-Based Systems, vol. 323 (113649). The Version of Record is available online at https://doi.org/10.1016/j.knosys.2025.113649., [Abstract]: We present a novel model-based time series clustering technique. Rather than fitting a model to each time series in isolation and then clustering the estimated model coefficients, our approach finds partitions where a single model accurately represents the entire group. This strategy exploits the inherent similarity between time series to get better model estimates, resulting in more robust and informative clusters. The models fitted to each partition deviate from the classic, ‘local’ time series model classes such as the ARIMA and instead follow the recent so-called ‘global’ forecasting models or cross-learning paradigm. Global models achieve superior predictive accuracy by fitting a more complex model class to the pool of time series in a dataset. This way, similarity information is better exploited and minor sources of heterogeneity are captured by the increased complexity. The procedure has a key additional main benefit. The problem of selecting the number of clusters, often separate from the partitioning process and subjective to the analyst, is completely solved in our case: the number of clusters that optimizes the predictive accuracy of the underlying global forecasting models should be chosen. In an extensive simulation and real-data study, we provide evidence of this approach outperforming reference techniques in both clustering quality and predictive accuracy. As the procedure is agnostic to the choice of the forecasting model, it can be combined with any model class. We provide examples of interpretation for linear models., The authors gratefully acknowledge the referees for their feedback on an earlier draft of this article. Ángel López-Oriona expresses his thanks to King Abdullah University of Science and Technology (KAUST) for its support. For José A. Vilar, this work is part of the grants PID2020-113578RB-I00 and PID2023-147127OB-I00 “ERDF/EU”, funded by MCIN/AEI/10.13039/501100011033/. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2024/14) and by CITIC as a center accredited for excellence within the Galician University System and a member of the CIGUS Network, receiving subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia. Additionally, it is co-financed by the European Union (EU) through the FEDER Galicia 2021–2027 operational program (Ref. ED431G 2023/01)., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Small area estimation of average compositions under multivariate nested error regression models

  • Esteban, M. Dolores
  • Lombardía, María José
  • López Vizcaíno, María Esther
  • Morales, Domingo
  • Pérez, Agustín
[Abstract]: This paper investigates the small area estimation of population averages of unit-level compositional data. The new methodology transforms the compositions into vectors of Rm and assumes that the vectors follow a multivariate nested error regression model. Empirical best predictors of domain indicators are derived from the fitted model, and their mean squared errors are estimated by parametric bootstrap. The empirical analysis of the behavior of the introduced predictors is investigated by means of simulation experiments. An application to real data from the Spanish household budget survey is given. The target is to estimate the average of proportions of annual household expenditures on food, housing and others, by Spanish provinces., Generalitat Valenciana; Prometeo/2021/063, Xunta de Galicia; ED431C 2020/14, Xunta de Galicia; ED431G 2019/01, Axencia Galega de Innovación; COV20/00604, Supported by the Instituto Galego de Estatística, by the Grants PGC2018-096840-B-I00 and PID2020-113578RB-I00 of the Spanish Ministerio de Economía y Competitividad, by the Grant Prometeo/2021/063 of the Generalitat Valenciana, and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C 2020/14), and by GAIN (Galician Innovation Agency) and the Regional Ministry of Economy, Employment and Industry Grant COV20/00604 and Centro de Investigación del Sistema Universitario de Galicia ED431G 2019/01, all of them through the ERDF.
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Nonparametric Kernel Estimation of the Probability of Cure in a Mixture Cure Model when the Cure Status is Partially Observed

  • Safari, Wende Clarence
  • López-de-Ullibarri, Ignacio
  • Jácome, M. A.
[Abstract] Cure models are a class of time-to-event models where a proportion of individuals will never experience the event of interest. The lifetimes of these so-called cured individuals are always censored. It is usually assumed that one never knows which censored observation is cured and which is uncured, so the cure status is unknown for censored times. In this paper, we develop a method to estimate the probability of cure in the mixture cure model when some censored individuals are known to be cured. A cure probability estimator that incorporates the cure status information is introduced. This estimator is shown to be strongly consistent and asymptotically normally distributed. Two alternative estimators are also presented. The first one considers a competing risks approach with two types of competing events, the event of interest and the cure. The second alternative estimator is based on the fact that the probability of cure can be written as the conditional mean of the cure status. Hence, nonparametric regression methods can be applied to estimate this conditional mean. However, the cure status remains unknown for some censored individuals. Consequently, the application of regression methods in this context requires handling missing data in the response variable (cure status). Simulations are performed to evaluate the finite sample performance of the estimators, and we apply them to the analysis of two datasets related to survival of breast cancer patients and length of hospital stay of COVID-19 patients requiring intensive care., The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research/work has been supported by MICINN grant PID2020-113578RB-I00, and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020/14 and Centro de Investigación del Sistema universitario de Galicia ED431G 2019/01), all of them through the ERDF, Xunta de Galicia; ED431C-2020/14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Estimation of patient flow in hospitals using up-to-date data. Application to bed demand prediction during pandemic waves

  • García-Vicuña, Daniel
  • López-Cheda, Ana
  • Jácome, M. A.
  • Mallor, Fermín
[Abstract] Hospital bed demand forecast is a first-order concern for public health action to avoid healthcare systems to be overwhelmed. Predictions are usually performed by estimating patients flow, that is, lengths of stay and branching probabilities. In most approaches in the literature, estimations rely on not updated published information or historical data. This may lead to unreliable estimates and biased forecasts during new or non-stationary situations. In this paper, we introduce a flexible adaptive procedure using only near-real-time information. Such method requires handling censored information from patients still in hospital. This approach allows the efficient estimation of the distributions of lengths of stay and probabilities used to represent the patient pathways. This is very relevant at the first stages of a pandemic, when there is much uncertainty and too few patients have completely observed pathways. Furthermore, the performance of the proposed method is assessed in an extensive simulation study in which the patient flow in a hospital during a pandemic wave is modelled. We further discuss the advantages and limitations of the method, as well as potential extensions., DGV and FM acknowledge the support by grant PID2020-114031RB-I00 (AEI, FEDER EU) and by the Government of Navarre, 0011-3597-2020-000003 (COVID). ALC was sponsored by the BEATRIZ GALINDO JUNIOR Spanish Grant from MICINN (Ministerio de Ciencia e Innovación) with code BGP18/00154. ALC and MAJ acknowledge partial support by the MICINN Grant PID2020-113578RB-I00 and partial support of Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14). ALC and MJ wish to acknowledge the support received from the Centro de Investigación de Galicia "CITIC", funded by Xunta de Galicia and the European Union European Regional Development Fund (ERDF)-Galicia 2014-2020 Program, by grant ED431G 2019/01. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript, Gobierno de Navarra; 0011-3597-2020-000003 (COVID), Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

On the Reliability of Machine Learning Models for Survival Analysis When Cure Is a Possibility

  • Ezquerro, Ana
  • Cancela, Brais
  • López-Cheda, Ana
[Abstract]: In classical survival analysis, it is assumed that all the individuals will experience the event of interest. However, if there is a proportion of subjects who will never experience the event, then a standard survival approach is not appropriate, and cure models should be considered instead. This paper deals with the problem of adapting a machine learning approach for classical survival analysis to a situation when cure (i.e., not suffering the event) is a possibility. Specifically, a brief review of cure models and recent machine learning methodologies is presented, and an adaptation of machine learning approaches to account for cured individuals is introduced. In order to validate the proposed methods, we present an extensive simulation study in which we compare the performance of the adapted machine learning algorithms with existing cure models. The results show the good behavior of the semiparametric or the nonparametric approaches, depending on the simulated scenario. The practical utility of the methodology is showcased through two real-world dataset illustrations. In the first one, the results show the gain of using the nonparametric mixture cure model approach. In the second example, the results show the poor performance of some machine learning methods for small sample sizes., This project was funded by the Xunta de Galicia (Axencia Galega de Innovación) Research projects COVID-19 presented in ISCIII IN845D 2020/26, Operational Program FEDER Galicia 2014–2020; by the Centro de Investigación de Galicia “CITIC”, funded by Xunta de Galicia and the European Union European Regional Development Fund (ERDF)-Galicia 2014–2020 Program, by grant ED431G 2019/01; and by the Spanish Ministerio de Economía y Competitividad (research projects PID2019-109238GB-C22 and PID2021-128045OA-I00). ALC was sponsored by the BEATRIZ GALINDO JUNIOR Spanish Grant from MICINN (Ministerio de Ciencia e Innovación) with code BGP18/00154. ALC was partially supported by the MICINN Grant PID2020-113578RB-I00 and partial support of Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14). We gratefully acknowledge the support of NVIDIA Corporation with the donation of the Titan Xp GPU used for this research., Xunta de Galicia; ED431G 2019/01, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; IN845D 2020/26
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Bootstrap Bandwidth Selection and Confidence Regions for Double Smoothed Default Probability Estimation

  • Peláez, Rebeca
  • Cao, Ricardo
  • Vilar, Juan M.
[Abstract] For a fixed time, t, and a horizon time, b, the probability of default (PD) measures the probability that an obligor, that has paid his/her credit until time t, runs into arrears not later that time t+b. This probability is one of the most crucial elements that influences the risk in credits. Previous works have proposed nonparametric estimators for the probability of default derived from Beran’s estimator and a doubly smoothed Beran’s estimator of the conditional survival function for censored data. They have also found asymptotic expressions for the bias and variance of the estimators, but they do not provide any practical way to choose the smoothing parameters involved. In this paper, resampling methods based on bootstrap techniques are proposed to approximate the bandwidths on which Beran and smoothed Beran’s estimators of the PD depend. Bootstrap algorithms for the calculation of confidence regions of the probability of default are also proposed. Extensive simulation studies show the good behavior of the presented algorithms. The bandwidth selector and the confidence region algorithm are applied to a German credit dataset to analyze the probability of default conditional on the credit scoring., This research has been supported by MICINN Grant PID2020-113578RB-100, and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro Singular de Investigación de Galicia ED431G 2019/01), all of them through the ERDF, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Poverty Mapping Under Area-Level Random Regression Coefficient Poisson Models

  • Diz-Rosales, Naomi
  • Lombardía, María José
  • Morales, Domingo
[Abstract]: Under an area-level random regression coefficient Poisson model, this article derives small area predictors of counts and proportions and introduces bootstrap estimators of the mean squared errors (MSEs). The maximum likelihood estimators of the model parameters and the mode predictors of the random effects are calculated by a Laplace approximation algorithm. Simulation experiments are implemented to investigate the behavior of the fitting algorithm, the predictors, and the MSE estimators with and without bias correction. The new statistical methodology is applied to data from the Spanish Living Conditions Survey. The target is to estimate the proportions of women and men under the poverty line by province., This work was supported by the Ministry of Science and Innovation and the State Research Agency of the Spanish Government through the European Regional Development Fund (PID2022-136878NB-I00, PID2020-113578RB-I00 and PRE2021-100857 to Naomi Diz-Rosales funded by MCIN/AEI/10.13039/501100011033); by the Conselleria d’Innovació, Universitats, Ciéncia i Societat Digital of the Generalitat Valenciana (Prometeo/2021/063); by the Consellería de Cultura, Educación, Formación Profesional e Universidades of the Xunta de Galicia through the European Regional Development Fund (Competitive Reference Groups ED431C/2020/14, COV20/00604, and ED431G/2019/01); and by Centro de Investigación en Tecnologías de la Información y las Comunicaciones (CITIC) that is supported by Xunta de Galicia, collaboration agreement between the Consellería de Cultura, Educación, Formación Profesional e Universidades and the Galician universities for the reinforcement of the research centers of the Sistema Universitario de Galicia (CIGUS)., Xunta de Galicia; ED431C/2020/14, Xunta de Galicia; COV20/00604, Xunta de Galicia; ED431G/2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Bagging cross-validated bandwidth selection in nonparametric regression estimation with applications to large-sized samples

  • Barreiro-Ures, Daniel
  • Cao, Ricardo
  • Francisco-Fernández, Mario
  • Fernández-Casal, Rubén
[Abstract]: Cross-validation is a well-known and widely used bandwidth selection method in nonparametric regression estimation. However, this technique has two remarkable drawbacks: the large variability of the selected bandwidths, and the inability to provide results in a reasonable time for very large sample sizes. To address these issues, bagged cross-validation bandwidth selectors are investigated. This approach consists in computing the cross-validation bandwidths for a finite number of subsamples and then rescaling the averaged smoothing parameters to the original sample size. Under a random-design regression model, asymptotic expressions up to a second-order for the bias and variance of the leave-one-out cross-validation bandwidth for the Nadaraya–Watson estimator are obtained. Subsequently, the asymptotic bias and variance and the limiting distribution for the bagged cross-validation selector are derived. Suitable choices of the number of subsamples and the subsample size lead to a convergence rate proportional to the inverse square root of the sample size for the bagging cross-validation selector, outperforming the slower rate typically associated with leave-one-out cross-validation. Several simulations and an illustration on a real dataset related to the COVID-19 pandemic show the behavior of our proposal and its better performance, in terms of statistical efficiency and computing time, when compared to leave-one-out cross-validation., This work is part of the grants PID2020-113578RB-I00 and PID2023-147127OB-I00 “ERDF/EU”, funded by MCIN/AEI/10.13039/501100011033/. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2024/14) and by CITIC as a center accredited for excellence within the Galician University System and a member of the CIGUS Network, receives subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia. Additionally, it is co-financed by the EU through the FEDER Galicia 2021-27 operational program (Ref. ED431G 2023/01). Funding for open access charge: Universidade da Coruña/CISUG., Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Semi-functional partial linear regression with measurement error: an approach based on kNN estimation

  • Novo Díaz, Silvia
  • Aneiros, Germán
  • Vieu, Philippe
This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: http://dx.doi.org/10.1007/s11749-024-00957-3, [Abstract]: This paper focuses on a semi-parametric regression model in which the response variable is explained by the sum of two components. One of them is parametric (linear), the corresponding explanatory variable is measured with additive error and its dimension is finite (p). The other component models, in a nonparametric way, the effect of a functional variable (infinite dimension) on the response. kNN-based estimators are proposed for each component, and some asymptotic results are obtained. A simulation study illustrates the behaviour of such estimators for finite sample sizes, while an application to real data shows the usefulness of our proposal., This research/work is part of the grants PID2020-113578RB-I00 and PID2023-147127OB-I00 ‘ERDF/EU’, funded by MCIN/AEI/10.13039/501100011033/. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2024/14) and by CITIC as a centre accredited for excellence within the Galician University System and a member of the CIGUS Network, receives subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia. Additionally, it is co-financed by the EU through the FEDER Galicia 2021-27 operational programme (Ref. ED431G 2023/01). The authors thank two anonymous referees for their constructive comments and suggestions which helped to improve the quality of this paper., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

On Functional Data Analysis and Related Topics

  • Aneiros, Germán
  • Horová, Ivana
  • Hušková, Marie
  • Vieu, Philippe
Accepted version: ©2022 Elsevier B.V. All rights reserved. This manuscript version is made available under the CC-BY-NC-ND 4.0 license https://creativecommons.org/licenses/by-nc-nd/4.0/. This version of the article has been accepted for publication in Journal of Multivariate Analysis. The Version of Record is available online at https://doi.org/10.1016/j.jmva.2021.104861, [Abstract]: This paper aims to present the various contributions to the Special Issue of the Journal of Multivariate Analysis on Functional Data Analysis and some related topics including High-Dimensional Statistics and Multivariate Analysis of complex data. The presentation is made by emphasizing on how the contributions are behaving among recent trends in the fields, in such a way that this paper can also be viewed as a selected bibliographical discussion about recent advances in these topics., The research by G. Aneiros and P. Vieu was supported in part by the MINECO (Ministerio de Economía, Industria 𝑦 Competitividad), Spain Grant MTM2017-82724-R and the MICINN (Ministerio de Ciencia e Innovación), Spain Grant PID2020-113578RB-I00 (EU ERDF support included). G. Aneiros also thanks the partial support by the Xunta de Galicia, Spain (Grupos de Referencia Competitiva ED431C-2020-14 and Centro Singular de Investigación del Sistema Universitario de Galicia ED431G 2019/01), all of them through the ERDF. The research of Marie Hušková was partially supported by Czech Science Foundation , grants GACR 18-08888S and 21-13323S., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01, Czech Republic. Czech Science Foundation; GACR 18-08888S, Czech Republic. Czech Science Foundation; 21-13323S
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Cost-sensitive reinforcement learning for credit risk

  • C-Rella, Jorge
  • Martínez Rego, David
  • Vilar, Juan M.
[Abstract]: Credit risk problems are dynamic because customer behavior is not stable, and they are cost-sensitive because the impact of a decision depends on the amount of the loan. Online learning algorithms, which evolve as more information becomes available, are an appropriate tool to study these dynamic problems. However, only information on approved transactions is available, which can lead to unfair biases and opportunity costs. Within reinforcement learning, bandit algorithms address this by balancing exploitation (acting according to the current model) and exploration (considering an action with limited information to improve predictions). The only remaining gap is to address the problem taking into account the different classification costs. This paper introduces cost-sensitive reinforcement learning algorithms to solve the credit risk problem from a dynamic perspective maximizing long-term benefits, proposing a cost-sensitive passive-aggressive algorithm and a cost-sensitive logistic bandit. Experiments on benchmark datasets and extensive simulation studies demonstrate the effectiveness and efficiency of the proposed algorithms, This research is part of the grants PID2020-113578RB-I00 and PID2023-147127OB-I00 "ERDF/EU", funded by MCIN/AEI/10.13039/ 501100011033/. It has also been supported by the Xunta de Galicia, Spain (Grupos de Referencia Competitiva ED431C-2024/14) and by CITIC as a center accredited for excellence within the Galician University System and a member of the CIGUS Network, receives subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia. Additionally, it is co-financed by the EU through the FEDER Galicia 2021–27 operational program (Ref. ED431G 2023/01). The first author was financed by the Axencia Galega de Innovación Industrial PhD Grant 14-IN606D-2021-2607768, Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01, Xunta de Galicia; 14-IN606D-2021-2607768
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Small area prediction of proportions and counts under a spatial Poisson mixed model

  • Boubeta, Miguel
  • Lombardía, María José
  • Morales, Domingo
[Abstract]: This paper introduces an area-level Poisson mixed model with SAR(1) spatially correlated random effects. Small area predictors of proportions and counts are derived from the new model and the corresponding mean squared errors are estimated by parametric bootstrap. The behaviour of the introduced predictors is empirically investigated by running model-based simulation experiments. An application to real data from the Spanish living conditions survey of Galicia (Spain) is given. The target is the estimation of domain proportions of women under the poverty line., Supported by the Instituto Galego de Estatística, by MICINN Grants PID2020-113578RB-I00 and PGC2018-096840-B-I00, by the Generalitat Valenciana Grant PROMETEO/2021/063 and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C 2020/14), and by GAIN (Galician Innovation Agency) and the Regional Ministry of Economy, Employment and Industry Grant COV20/00604 and Centro de Investigación del Sistema Universitario de Galicia ED431G 2019/01, all of them through the ERDF., Generalitat Valenciana; PROMETEO/2021/063, Xunta de Galicia; ED431C/2020/14, Xunta de Galicia; COV20/00604, Xunta de Galicia; ED431G/2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Tesis doctoral (DoctoralThesis)

High dimensional single-index mixture cure models

  • Piñeiro Lamas, Beatriz
[Abstract] In survival analysis, there are situations in which not all subjects are susceptible to the final event. For example, if the event is a cancer therapy-related adverse effect, there will be a fraction of patients (considered as cured) that will never experience it. Mixture cure models allow to estimate the probability of cure and the survival function for the uncured subjects. In the literature, nonparametric estimation of both functions is limited to continuous univariate covariates. We fill this gap by proposing single-index mixture cure models. They allow working with a vector covariate and assume that the survival function depends on it through an unknown linear combination, that is estimated by maximum likelihood. Besides, a nonparametric estimator for the density function of the uncured individuals is introduced, and its iid representation is derived. Finally, the proposed models are extended to functional covariates and a preprocessing algorithm is implemented to deal with medical images. The methodology is applied to a cardiotoxicity dataset. The goal is to determine whether (and how) certain factors affect the probability of experiencing the cardiovascular problem and the amount of time it takes for it to manifest. Understanding risk factors may lead to a patient-based preventive medicine., [Resumen] En análisis de supervivencia hay situaciones en las cuales no todos los sujetos son susceptibles al evento final. Por ejemplo, si el evento es un efecto adverso de una terapia oncológica, habrá pacientes (considerados como curados) que nunca lo experimentarán. Los modelos de curación de tipo mixtura permiten estimar la probabilidad de curación y la función de supervivencia de los sujetos susceptibles. En la literatura, la estimaciuñin no paramétrica de ambas funciones se limita a covariables unidimensionales. En esta tesis se proponen los modelos single-index de curación de tipo mixtura. Estos permiten trabajar con una covariable vectorial y asumen que la función de supervivencia depende de ella mediante una combinación lineal que se puede estimar por máxima verosimilitud. Además, se introduce un estimador no paramétrico para la funci´on de densidad de los susceptibles y se obtiene su representación iid. Finalmente, los modelos propuestos se extienden a covariables funcionales y se implementa un algoritmo de preprocesamiento de imágenes médicas. La metodología se aplica a datos de cardiotoxicidad con el objetivo de determinar qué factores afectan (y cómo) a la probabilidad de experimentarla y al tiempo que tarda en manifestarse. Conocer los factores de riesgo podría conducir a una medicina preventiva personalizada., [Resumo] En análise de supervivencia hai situacións nas que non todos os individuos son susceptibles ao evento final. Por exemplo, se o evento é un efecto adverso dunha terapiaoncolóxica, haberá pacientes (considerados como curados) que nunca o experimentarán. Os modelos de curaci´on de tipo mestura permiten estimar a probabilidade de curación e a función de supervivencia dos suxeitos susceptibles. Na literatura, a estimación non paramétrica destas dúas funcións limítase a covariables unidimensionais. Nesta tese propóñense os modelos single-index de curación de tipo mestura. Estes permiten traballar cunha covariable vectorial e asumen que a función de supervivencia depende dela mediante unha combinación lineal que se pode estimar por máxima verosimilitude. Ademais, introd´ucese un estimador non paramétrico para a función de densidade dos susceptibles e obtense a súa representación iid. Finalmente, os modelos propostos esténdense a covariables funcionais e impleméntase un algoritmo de preprocesamento de imaxes médicas. A metodoloxía aplícase a datos de cardiotoxicidade co obxectivo de determinar que factores afectan (e como) á probabilidade de experimentala e ao tempo que tarda en manifestarse. Coñecer os factores de risco podería conducir cara unha medicina preventiva personalizada., Xunta de Galicia; ED431C 2020/14, Xunta de Galicia; ED431G 2019/01, Xunta de Galicia; ED481A-2020/290, This research has been partially supported by MINECO (Ministerio de Economía y Competitividad) Grant MTM2017-82724-R, by Ministerio de Ciencia e Innovación under grant PID2020-113578RB-100 and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C 2020/14 and Centro de Investigación del Sistema Universitario de Galicia ED431G 2019/01), all of them through the ERDF (European Regional Development Fund). The doctoral student acknowledges financial support from Axudas Predoutorais da Xunta de Galicia, with reference ED481A-2020/290. The work has been partially carried out during two visits at École Nationale de la Statistique et de l’Analyse de l’Information (ENSAI) and Katholieke Universiteit Leuven. The first stay was funded through grant ED481A-2020/290 and the second one was supported by the research group MODES (Modelización, Optimización e Inferencia Estadística) and CITIC (Centro de Investigación de Tecnologías de la Información y de la Comunicación).
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Cost-sensitive single-index classification model

  • C-Rella, Jorge
  • Cao, Ricardo
  • Vilar, Juan M.
[Abstract]: Single-index models (SIMs) are a type of semiparametric model in which a response variable is assumed to be related to a linear combination of explanatory variables by an unknown function, on which any restriction is imposed. Thus, they provide both interpretability and flexibility to capture complex data relationships. In this paper, SIMs are extended to the cost-sensitive classification problem by minimizing the different misclassification costs. The flexibility of SIMs combined with a cost-sensitive approach results in a powerful model to minimize losses and optimize decision making. This is demonstrated through an extensive simulation study and the analysis of five real data sets, where the proposed approach outperforms both parametric and semi-parametric previous approaches., This research is part of the grants PID2020-113578RB-I00 and PID2023-147127OB-I00 “ERDF/EU”, funded by MCIN/AEI/10.13039/ 501100011033/. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2024/14) and by CITIC as a center accredited for excellence within the Galician University System and a member of the CIGUS Network, receives subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia, Spain. Additionally, it is co-financed by the EU through the FEDER Galicia 2021-27 operational program (Ref. ED431G 2023/01). The first author was financed by the Axencia Galega de Innovación, Spain Grant 14-IN606D-2021-2607768., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01, Xunta de Galicia; 14-IN606D-2021-2607768
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Robust Methods for Soft Clustering of Multidimensional Time Series

  • López-Oriona, Ángel
  • D'Urso, Pierpaolo
  • Vilar, José
  • Lafuente-Rego, Borja
Presented at the 4th XoveTIC Conference, A Coruña, Spain, 7–8 October 2021., [Abstract] Three robust algorithms for clustering multidimensional time series from the perspective of underlying processes are proposed. The methods are robust extensions of a fuzzy C-means model based on estimates of the quantile cross-spectral density. Robustness to the presence of anomalous elements is achieved by using the so-called metric, noise and trimmed approaches. Analyses from a wide simulation study indicate that the algorithms are substantially effective in coping with the presence of outlying series, clearly outperforming alternative procedures. The usefulness of the suggested methods is also highlighted by means of a specific application., This research has been supported by MINECO (MTM2017-82724-R and PID2020-113578RB-100), the Xunta de Galicia (ED431C-2020-14), and “CITIC” (ED431G 2019/01)., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Bootstrap-LOCI data mining methodology for anomaly detection in buildings energy efficiency

  • Tobar, Andrés
  • Flores Sánchez, Miguel
  • Castillo-Páez, Sergio
  • Naya, Salvador
  • Zaragoza, Sonia
  • Tarrío-Saavedra, Javier
[Abtract]: An automated methodology is proposed to identify anomalies in buildings’ HVAC systems, through Local Correlation Integral (LOCI) algorithm, improved by Bootstrap to obtain a rule from its score distribution. It has been performed to solve the case study of anomaly detection for HVAC facilities maintenance in a clothing store in Panama. It is defined by a dataset composed of 24 daily readings recorded during 434 days. In each reading, 15 critical quality variables for thermal comfort and energy efficiency were monitored. For algorithm training, anomalous events recorded by HVAC system operators are considered. In this stage, the LOCI parameters that best fit the data are estimated, to obtain a score for each of the observations and then study their distribution by applying Bootstrap techniques to improve the classification performance. For the algorithm performance evaluation, crossvalidation is used and from these results, it is compared with two benchmark supervised classification methods such as Logistic Regression and Support Vector Machines with polynomial kernel. The LOCI method has been improved by the application of bootstrap resampling, providing estimates of the LOCI score distribution and a critical value from which we define an anomalous observation. It provides the best balance between real anomaly detection and prevention of false alarm identification than current proposals for LOCI., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G2019/01, The research has been supported by Ministerio de Ciencia e Innovación, Spain grant PID2020-113578RB-100, and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro de Investigación del Sistema universitario de Galicia ED431G2019/01), all of them through the ERDF.
RUC. Repositorio da Universidade da Coruña
Publicación Tesis doctoral (DoctoralThesis)

Methodological Contributions in Semiparametric Regression Models for Functional Data

  • Novo Díaz, Silvia
[Abstract] This doctoral thesis is dedicated to functional regression for scalar response. In particular, we focus on functional semiparametric models, which combine the practical advantages of parametric and nonparametric approaches, surpassing both methodologies. Accordingly, several semiparametric models involving a functional singleindex component were studied from a theoretical and practical perspective. First, for the functional single-index model (FSIM) and the semi-functional partial linear single-index model (SFPLSIM), we provide uniform consistency results (over all parameters involved) for kernel- and k-Nearest-Neighbours-based statistics related to the estimation of the semiparametric component. Second, for the sparse semifunctional partial linear single-index model (SSFPLSIM), we develop a variable selection procedure in the linear component based on penalized least squares (PLS). The good behaviour of this method is theoretically assured (rates of convergence of the estimators are obtained, as well as asymptotic behaviour of the variable selection procedure). Third, the SSFPLSIM is adapted to the case in which covariates with linear e ect come from the discretization of a curve. For this new model, the multi-functional partial linear single-index model (MFPLSIM), the variable selection problem was also studied. Consequently, two new algorithms were proposed (providing theoretical results that ensure their good performance) to solve the ine ciency of the PLS method when it is directly applied to the MFPLSIM. For all the models and procedures mentioned above, theoretical results are accompanied by both simulation studies and real data applications which illustrate the good performance of the proposed methodology in practice., [Resumo] Esta tese está adicada ao estudo da regresión funcional con variable resposta escalar. En particular, centrámonos en modelos funcionais semi-paramétricos, os cales combinan as vantaxes prácticas dos enfoques paramétrico e non-paramétrico, superando a ambas metodoloxías. Desta maneira, estudáronse, tanto dende o punto de vista te orico como dende a perspectiva pr actica, varios modelos semi-param etricos que involucran unha compoñente funcional single-index. En primeiro lugar, para o functional single-index model (FSIM) e para o semi-functional partial linear single-index model (SFPLSIM) establecemos resultados de consistencia uniforme (sobre todos os parámetros involucrados) para os estatísticos de tipo núcleo e tipo k-veciños-máis-próximos relacionados coa estimación da compoñente semi-paramétrica do modelo. En segundo lugar, para o sparse semi-functional partial linear single-index model (SSFPLSIM) desenvolvemos un procedemento de selección de variables na compoñente linear baseado en mínimos cadrados penalizados (PLS, iniciais de penalized least squares). O bo comportamento deste método asegurouse dende o punto de vista teórico (obtendo taxas de converxencia dos estimadores, así como o comportamento asintótico do procedemento de selección de variables). En terceiro lugar, o SSFPLSIM adaptouse ao escenario no cal as covariables con efecto linear proveñen da discretización dunha curva. Para este novo modelo, o multi-functional partial linear single-index model (MFPLSIM), estudouse tamén o problema da selección de variables e propuxéronse dous novos algoritmos (dos que aseguramos teoricamente o seu bo comportamento) para resolver a inefi cacia do método PLS cando se aplica directamente ao MFPLSIM. Para todos os modelos e procedementos citados, os resultados teóricos acompañáronse de estudos de simulación e aplicacións a datos reais que ilustran o bo comportamento na práctica da metodoloxía presentada., [Resumen] Esta tesis está dedicada al estudio de la regresión funcional con variable respuesta escalar. En particular, nos centramos en modelos funcionales semi-paramétricos, los cuales combinan las ventajas prácticas de los enfoques paramétrico y no-paramétrico, superando a ambas metodologías. De esta forma, se estudiaron, tanto desde el punto de vista teórico como desde la perspectiva práctica, varios modelos semi-paramétricos que involucran una componente funcional single-index. En primer lugar, para el functional single-index model (FSIM) y para el semi-functional partial linear single-index model (SFPLSIM) establecemos resultados de consistencia uniforme (sobre todos los parámetros involucrados) para los estadísticos de tipo núcleo y de tipo k-vecinos-más-próximos relacionados con la estimación de la componente semi-paramétrica del modelo. En segundo lugar, para el sparse semi-functional partial linear single-index model (SSFPLSIM) desarrollamos un procedimiento de selección de variables en la componente linear basado en mínimos cuadrados penalizados (PLS, iniciales de penalized least squares). El buen comportamiento de este método se ha asegurado desde el punto de vista teórico (obteniendo tasas de convergencia de los estimadores, así como el comportamiento asintótico del procedimiento de selección de variables). En tercer lugar, el SSFPLSIM se ha adaptado al escenario en el cual las covariables con efecto linear provienen de la discretización de una curva. Para este nuevo modelo, el multi-functional partial linear single-index model (MFPLSIM), se ha estudiado también el problema de selección de variables y se propusieron dos nuevos algoritmos (de los que aseguramos teóricamente su buen comportamiento) para resolver la ine ficiencia del método PLS cuando se aplica directamente al MFPLSIM. Para todos los modelos y procedimientos citados, los resultados teóricos se acompañaron de estudios de simulación y aplicaciones a datos reales que ilustran el buen comportamiento en la práctica de la metodología presentada., Xunta de Galicia; ED481A-2018/191, Xunta de Galicia; ED431G/01 2016-2019, Xunta de Galicia; ED431G2019/01, Xunta de Galicia; ED431C 2016-015, Xunta de Galicia; ED431C2020-014, This research has been partially supported by the Spanish Ministerio de Economía y Competitividad (MINECO) under Grants MTM2014-52876-R and MTM2017-82724-R, by the Spanish Ministerio de Ciencia e Innovación (MICINN) under Grant PID2020-113578RB-I00, by the Xunta de Galicia through Centro Singular de Investigación de Galicia accreditation under Grants ED431G/01 2016-2019 and ED431G 2019/01 and through the Grupos de Referencia Competitiva under Grants ED431C 2016-015 and ED431C2020-014 and in part by the European Union (European Regional Development Fund-ERDF). The author particularly thanks the contracts financed by the research group MODES (from May 1, 2017 to August 31, 2017) and by the CITIC (from September 1, 2017 to May 30, 2018) and the PhD contract financed by the Xunta de Galicia and the European Union (European Social Fund-ESF), the reference of which is ED481A-2018/191 (from May 31, 2018). Some results of this thesis have been obtained during a stay of the author at the Universit ́e Paul Sabatier, Toulouse (from March 13, 2019 to June 12, 2019), financed by the Xunta de Galicia, with reference ED481A-2018/191
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

CABRA: Clustering algorithm based on regular arrangement

  • C-Rella, Jorge
[Abstract]: Clustering is an unsupervised learning technique for organizing complex datasets into coherent groups. A novel clustering algorithm is presented, with a simple grouping concept depending on only one hyperparameter, which makes it suitable for further extensions to any topology and space. It is compared to state-of-the-art algorithms, overall achieving a better performance independently on the structure and complexity of the data, making the proposed algorithm a valuable tool for real applications such as market segmentation, sentiment analysis and anomaly detection., This research has been financed by the Grant PID2020-113578RB-I00, funded by MCIN/AEI/10.2039/501100011033/. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020/14) and by CITIC that is supported by Xunta de Galicia, convenio de colaboración entre la Consellería de Cultura, Educación, Formación Profesional e Universidades y las universidades gallegas para el refuerzo de los centros de investigación del Sistema Universitario de Galicia (CIGUS). The first author was financed by the Axencia Galega de Innovación Industrial PhD Grant 14-IN606D-2021-2607768., Xunta de Galicia; ED431C-2020/14, Xunta de Galicia; 14-IN606D-2021-2607768
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Erosive Degradation Study of Concrete Augmented by Mussel Shells for Marine Construction

  • Camba, C.
  • Mier, Jose
  • Carral Couce, Luis
  • Lamas, M.I.
  • Álvarez-Feal, José-Carlos
  • Díaz-Díaz, Ana-María
  • Tarrío-Saavedra, Javier
[Abstract] This work proposes a green material for artificial reefs to be placed in Galicia (northwest Spain) taking into account the principles of circular economy and sustainability of the ecosystem. New concrete formulations for marine applications, based on cement and/or sand replacement by mussel shells, are analyzed in terms of resistance to abrasion. The interest lies in the importance of the canning industry of Galicia, which generates important quantities of shell residues with negative environmental consequences. Currently, the tests to determine the abrasion erosion resistance of concrete on hydraulic structures involve large and complex devices. According to this, an experimental test has been proposed to estimate and compare the wear resistance of these concretes and, consequently, to analyze the environmental performance of these structures. First, a numerical analysis validated with experimental data was conducted to design the test. Subsequently, experimental tests were performed using a slurry tank in which samples with conventional cement and sand were partially replaced by mussel shell. The abrasive erosion effect of concrete components was analyzed by monitoring the mass loss. It shows an asymptotic trend with respect to time that has been modeled by Generalized Additive Model (GAM) and nonlinear regression models. The results were compared to concrete containing only conventional cement and sand. Replacing sand and/or cement by different proportions of mussel shells has not significantly reduced the resistance of concrete against erosive degradation, except for the case where a high amount of sand (20 wt.%) is replaced. Its resistance against the erosive abrasion is increased, losing between 0.1072 and 0.0310 wt.% lower than common concrete. In all the remaining cases (replacements of the 5–10 wt.% of sand and cement), the effect of mussel replacement on erosive degradation is not significant. These results encourage the use of mussel shells in the composition of concrete, taking into account that we obtain the same degradation properties, even more so considering an important residue in the canning industry (and part of the seabed) that can be valorized., This research was funded by Xunta de Galicia through the project CN-10MMA003CT. This study was also funded through the collaboration agreement between Xunta de Galicia (Convenio de colaboración entre a Xunta de Galicia, a Universidade da Coruña e a Fundación da Universidade da Coruña para dar continuidade ao Proxecto de investigación de arrecife artificial—PROARR), Universidade da Coruña and the Universidade da Coruña Foundation (FUAC) to give continuity to the previous project. The research of Javier Tarrío-Saavedra has been supported by MINECO grant MTM2017-82724-R, Ministerio de Ciencia e Innovación grant PID2020-113578RB-100, and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro de Investigación del Sistema universitario de Galicia ED431G2019/01), all of them through the ERDF, Xunta de Galicia; 10MMA003CT, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Ordinal Time Series Analysis with the R Package otsfeatures

  • López-Oriona, Ángel
  • Vilar, José
[Abstract] The 21st century has witnessed a growing interest in the analysis of time series data. While most of the literature on the topic deals with real-valued time series, ordinal time series have typically received much less attention. However, the development of specific analytical tools for the latter objects has substantially increased in recent years. The R package otsfeatures attempts to provide a set of simple functions for analyzing ordinal time series. In particular, several commands allowing the extraction of well-known statistical features and the execution of inferential tasks are available for the user. The output of several functions can be employed to perform traditional machine learning tasks including clustering, classification, or outlier detection. otsfeatures also incorporates two datasets of financial time series which were used in the literature for clustering purposes, as well as three interesting synthetic databases. The main properties of the package are described and its use is illustrated through several examples. Researchers from a broad variety of disciplines could benefit from the powerful tools provided by otsfeatures., This research has been supported by the Ministerio de Economía y Competitividad (MINECO) grants MTM2017-82724-R and PID2020-113578RB-100, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14), and the Centro de Investigación del Sistema Universitariode Galicia, “CITIC” grant ED431G 2019/01; all of them through the European Regional Development Fund (ERDF)., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Analyzing categorical time series with the R package ctsfeatures

  • López-Oriona, Ángel
  • Vilar, José
[Absctract]: Time series data are ubiquitous nowadays. Whereas most of the literature on the topic deals with real-valued time series, categorical time series have received much less attention. However, the development of data mining techniques for this kind of data has substantially increased in recent years. The R package ctsfeatures offers users a set of useful tools for analyzing categorical time series. In particular, several functions allowing the extraction of well-known statistical features and the construction of illustrative graphs describing underlying temporal patterns are provided in the package. The output of some functions can be employed to perform traditional machine learning tasks including clustering, classification and outlier detection. The package also includes two datasets of biological sequences introduced in the literature for clustering purposes, one dataset of sleep stages, and three interesting synthetic databases. In this work, the main characteristics of the package are described and its use is illustrated through various examples. Practitioners from a wide variety of fields could benefit from the valuable tools provided by ctsfeatures., The authors are grateful to the anonymous referees for their comments and suggestions. This research has been supported by the Ministerio de Economía y Competitividad (MINECO) grants MTM2017-82724-R and PID2020-113578RB-100, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14), and the Centro de Investigación del Sistema Universitario de Galicia, “CITIC” grant ED431G 2019/01; all of them through the European Regional Development Fund (ERDF) . This work has received funding for open access charge by Universidade da Coruña/CISUG., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Control y detección de anomalías de confort térmico y eficiencia energética en edificios a partir de variables autocorreladas

  • Barbeito, Inés
  • Zaragoza, Sonia
  • Tarrío-Saavedra, Javier
  • Naya, Salvador
[Resumen]: En este trabajo se presenta, de una forma intuitiva, paso por paso, la metodología propuesta en el artículo “Assessing thermal comfort and energy efficiency in buildings by statistical quality control for autocorrelated data” para el control de procesos definidos por una variable (critical to quality variable, CTQ) cuyas observaciones están autocorreladas, permitiendo, a su vez, la detección de anomalías, siendo éste su principal objetivo. De hecho, el origen de este procedimiento estáa en la necesidad de detección automática de alarmas que tienen las empresas relacionadas con las instalaciones de energía en edificios, en el marco de la Industria 4.0 y la digitalización de la producción y los servicios. En concreto, la empresa Σqus, desarrolladora de plataformas de Internet of Things (IoT) en el sector de la energía en edificaciones (centros comerciales, oficinas, hoteles, hospitales) presentó a los autores su necesidad de implementar algoritmos que permitieran la detección automática de anomalías en sus sistemas de climatización (Heating, Ventilation, and Air Conditioning, HVAC) a partir de variables como la temperatura ambiente interior o el consumo de energía en climatización, es decir, anomalías relacionadas con la eficiencia energética y de confort térmico del edificio. En consecuencia, se propuso la aplicación de gráficos de control x, de medidas individuales y de medias ponderadas exponencialmente (EWMA) a las variables CTQ. Dado que sus observaciones no son independientes (como es común en datos monitorizados continuamente con respecto al tiempo), se propone un procedimiento para aplicar estos gráficos a los residuos correspondientes al previo ajuste de modelos ARIMA para series de tiempo. El estudio se complementa con el análisis de la capacidad de las instalaciones HVAC del edificio para cumplir las especificaciones en términos de consumo energía y confort térmico, a través de la propuesta de aplicación de índices de capacidad univariante para datos autocorrelados., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G2019/01, Este estudio ha sido apoyado por el Ministerio de Ciencia e Innovación con la subvención PID 2020-113578RB-100, además de por la Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 y Centro de Investigación del Sistema universitario de Galicia ED431G2019/01), todos ellos a través de los fondos FEDER. Los autores quisieran agradecer a la SEIO su apoyo e interés en nuestro trabajo.
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Influence of Storage Time on the DNA Integrity and Viability of Spermatozoa of the Spider Crab Maja brachydactyla

  • Rodríguez Pena, Elba
  • Suárez, Diego
  • Estévez-Pérez, G.
  • Verísimo Amor, Patricia
  • Barreira, Noelia
  • Fernández, Luis
  • González-Tizón, Ana M.
  • Martínez-Lage, Andrés
The GTK application for comet segmentation developed in this work is available at https://github.com/dsuarezgarcia/freecomet. The rest of the data presented in this study are available on request from the corresponding author., [Abstract] Natural populations of the spider crab Maja brachydactyla constitute a fishery resource of great economic importance in many countries. As in the rest of eubrachyurans, the females of this species have ventral-type seminal receptacles where they store sperm from copulations. Sperm can be stored in these structures for months and even years before egg fertilisation, with the consequent degradation of the sperm cells during the time. In this work, we analyse the viability and the possible genetic damage in sperm accumulated in the seminal receptacles of M. brachydactyla females as a function of the storage time (from 0 to 14 months) using the comet assay technique. On one hand, we developed an algorithm for comet image analysis that improves the comet segmentation compared with the free software Open comet v1.3.1 (97% vs. 76% of detection). In addition, our software allows the manual modification of the contours wrongly delimited via the automatic tool. On the other hand, our data show a sharp decline in sperm viability and DNA integrity in the first four months of storage, which could lead to a decrease in the fecundity rate and/or viability of the embryos or larvae from the second and third clutches of the annual cycle if the repair capacity in these gametic cells is low., This research was funded by the Ministerio de Ciencia, Innovación y Universidades, grant number PID2019-109194RB-I00 MAR; and by the Ministerio de Ciencia e Innovación, grant number PID2020-113578RB-I00. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020/14)., Xunta de Galicia; ED431C-2020/14
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

TTS package: Computational tools for the application of the Time Temperature Superposition principle

  • Meneses Freire, Antonio
  • Naya, Salvador
  • Francisco-Fernández, Mario
  • López-Beceiro, Jorge
  • Gracia-Fernández, Carlos
  • Tarrío-Saavedra, Javier
[Abstract]: The TTS package has been developed in R software to predict the mechanical properties of viscoelastic materials, at short and long observation times/frequencies by applying the Time Temperature Superposition (TTS) principle. TTS is a physical principle used in material science to estimate mechanical properties beyond the experimental range of observed times/frequencies by shifting data curves obtained at other temperatures relative to a reference temperature in the dataset. It is a methodology related to accelerated life-tests and reliability, whereas the TTS library is one of the first open source computational tool to apply the TTS principle. This R package provides free computational tools to obtain master curves that characterize materials from a thermal-mechanical approach. The TTS package also proposes, implements and explains our own method to obtain the shift factors and the master curve in a TTS analysis, based on horizontal shifting of the first derivative function of viscoelastic properties. This procedure provides shift factors estimates and smooth master curve estimates using B-spline fitting, in a fully automatic way, without assuming any parametric expression. Williams-Landel-Ferry (WLF) and Arrhenius TTS parametric models are also implemented in the TTS package. They can be fitted from shifts obtained by the our first derivative based method., This work has been supported by Campus Industrial de Ferrol (Universidade da Coruña), Ministerio de Ciencia e Innovación grant PID2020-113578RB-100 and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2016-015 and Centro Singular de Investigación de Galicia ED431G/01), all of them through the ERDF. The authors would like to express their gratitude to the two reviewers of this article, whose contributions and suggestions have made a significant contribution to providing users with a truly useful tool., Xunta de Galicia; ED431C-2016-015, Xunta de Galicia; ED431G/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Goodness-of-fit tests for multiple regression with circular response

  • Meilán-Vila, Andrea
  • Francisco-Fernández, Mario
  • Crujeiras-Casais, Rosa M.
Versión final aceptada de: https://doi.org/10.1080/00949655.2021.2015597, This is an Accepted Manuscript of an article published by Taylor & Francis in Journal of Statistical Computation and Simulation on 2022, available at: https://doi.org/10.1080/00949655.2021.2015597, [Abstract]: Testing procedures for assessing a parametric regression model with a circular response and an Rd-valued covariate are proposed and analysed in this work. The test statistics are based on a circular distance comparing a (non-smoothed or smoothed) parametric circular regression estimator and a nonparametric one. Two bootstrap procedures for calibrating the tests in practice are also presented. Finite sample performance of the tests in different scenarios is analysed by simulations and illustrated with real data examples., The authors thank Prof. Felicita Scapini and her research team who kindly provided the sand hoppers data that are used in this work. Data were collected within the Project ERB ICI8-CT98-0270 from the European Commission, Directorate General XII Science. The authors also thank the Associate Editor and two anonymous referees for numerous useful comments that significantly improved this article. This research has been partially supported by MINECO (Grants MTM2016-76969-P and MTM2017-82724-R), MICINN (Grant PID2020-113578RB-I00) and by Xunta de Galicia (Grant ED481A-2017/361, through the ESF. Grupos de Referencia Competitiva ED431C-2016-015, ED431C-2017-38 and ED431C-2020-14, and Centro de Investigación del SUG ED431G 2019/01, through the ERDF)., Xunta de Galicia; ED481A-2017/361, Xunta de Galicia; ED431C-2016-015, Xunta de Galicia; ED431C-2017-38, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Nonparametric kernel estimation of cure models in survival analysis

  • López-Cheda, Ana
  • Cao, Ricardo
  • Jácome, M. A.
  • Keilegom, Ingrid Van
[Abstract]: In lifetime data, there may be long-term survivors, which lead to heavy censoring at the end of the follow-up period. In this context, a standard survival model is not appropriate, and therefore a cure model is needed. In the literature, the covariate effect in mixture cure models was modeled using parametric or semiparametric approaches. Recently, completely nonparametric methods for the cure rate and the latency were proposed in mixture cure models. In this paper, we present this methodology and apply it to a colorectal cancer dataset., The first three authors’ research has been partially supported by the grants PID2020-113578RB-I00 and PID2023-147127OB-I00 ”ERDF/EU”, funded by MCIN/AEI/10.13039/501100011033/. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2024/14) and by CITIC as a center accredited for excellence within the Galician University System and a member of the CIGUS Network, receives subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia. Additionally, it is co-financed by the EU through the FEDER Galicia 2021-27 operational program (Ref. ED431G 2023/01). Besides, this work integrated into the framework of PERTE for Cutting-edge Health, has been co-financed by the Spanish Ministry of Science, Innovation and Universities with funds from the European Union NextGenerationEU, from the Recovery, Transformation and Resilience Plan (PRTR-C17.I1) and from the Autonomous Community of Galicia within the framework of the Biotechnology Plan Applied to Health. Ingrid Van Keilegom gratefully acknowledges funding from the FWO and F.R.S. - FNRS (Excellence of Science programme, project ASTeRISK, grant no. G0I3422N), and from the FWO (senior research projects fundamental research, grant no. G047524N)., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01, Bélgica. Fonds Wetenschappelijk Onderzoek; G0I3422N, Bélgica. Fonds Wetenschappelijk Onderzoek; G047524N, Bélgica. Fonds de la Recherche Scientifique; G047524N
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Probability of default estimation in credit risk using mixture cure models

  • Peláez, Rebeca
  • Keilegom, Ingrid Van
  • Cao, Ricardo
  • Vilar, Juan M.
[Abstract]: An estimator of the probability of default (PD) in credit risk is proposed. It is derived from a nonparametric conditional survival function estimator based on cure models. Asymptotic expressions for the bias and the variance, as well as the asymptotic normality of the proposed estimator are presented. A simulation study shows the performance of the nonparametric estimator compared with Beran's PD estimator and other semiparametric methods. Finally, an empirical study based on modified real data illustrates the practical behaviour., This research has been supported by MICINN Grant PID2020-113578RB-100 , by the Xunta de Galicia (Grupo de Referencia Competitiva ED431C-2020-14 and Centro Singular de Investigación de Galicia ED431G 2019/01), all of them through the ERDF and by the European Research Council (2016-2022, Horizon 2020 / ERC grant agreement No. 694409 ). Peláez, R. was sponsored by inMOTION Programme of grants for pre-doctoral stays Inditex-UDC 2021., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Tesis doctoral (DoctoralThesis)

Statistical Inference in Mixed Models With Random Slope: Applications of Interest In Small Areas

  • Diz Rosales, Naomi
[Resumen] La estimación en áreas pequeñas aborda la obtención de estimaciones fiables de indicadores cuando los tamaños muestrales a nivel de dominio son reducidos o inexistentes. En este contexto, los modelos mixtos constituyen una herramienta fundamental al permitir combinar información global y específica de cada área. Sin embargo, la mayoría de los enfoques existentes se basan en modelos con interceptos aleatorios, asumiendo homogeneidad en los efectos de las variables explicativas entre dominios, una hipótesis a menudo demasiado restrictiva en aplicaciones reales. Esta tesis desarrolla un marco metodológico unificado para la inferencia en modelos de estimación en áreas pequeñas que incorporan pendientes aleatorias, permitiendo modelar explícitamente la heterogeneidad estructural entre áreas. Se estudian modelos de área con respuesta discreta (Poisson) y continua (Fay-Herriot), tanto en contextos transversales como longitudinales. Para cada modelo se derivan predictores óptimos, se proponen procedimientos eficientes de estimación de parámetros y se desarrollan estimadores del error cuadrático medio, tanto analíticos como basados en bootstrap. El comportamiento de las metodologías propuestas se evalúa mediante estudios de simulación y se ilustran sus ventajas en aplicaciones reales, incluyendo la estimación de ocupación hospitalaria en contextos epidémicos y la obtención de indicadores socioeconómicos desagregados a partir de encuestas oficiales. En conjunto, la tesis amplía el marco metodológico de la estimación en áreas pequeñas, proporcionando modelos más flexibles y realistas para la producción de indicadores fiables en situaciones de escasez de información., [Resumo] A estimación en áreas pequenas ocúpase da obtención de estimacións fiables de indicadores cando os tamaños mostrais a nivel de dominio son reducidos ou mesmo inexistentes. Neste contexto, os modelos mixtos constitúen unha ferramenta fundamental ao permitir combinar información global coa específica de cada área. Porén, a maioría dos enfoques existentes baséanse en modelos con intersectos aleatorios, asumindo homoxeneidade nos efectos das variables explicativas entre dominios, unha hipótese que adoita resultar excesivamente restritiva en aplicacións reais. Esta tese desenvolve un marco metodolóxico unificado para a inferencia en modelos de estimación en áreas pequenas que incorporan pendentes aleatorias, permitindo modelar explicitamente a heteroxeneidade estrutural entre áreas. Estúdanse modelos de área con resposta discreta (Poisson) e continua (Fay-Herriot), tanto en contextos transversais como lonxitudinais. Para cada modelo derívanse preditores óptimos, propóñense procedementos eficientes para a estimación dos parámetros e desenvólvense estimadores do erro cadrático medio, tanto analíticos como baseados en bootstrap. O comportamento das metodoloxías propostas avalíase mediante estudos de simulación e ilústranse as súas vantaxes en aplicacións reais, incluíndo a estimación da ocupación hospitalaria en contextos epidémicos e a obtención de indicadores socioeconómicos desagregados a partir de enquisas oficiais. En conxunto, esta tese amplía o marco metodolóxico da estimación en áreas pequenas, proporcionando modelos máis flexibles e realistas para a produción de indicadores fiables en situacións de escaseza de información., [Abstract] Small area estimation focuses on producing reliable estimates of indicators when sample sizes at the domain level are small or even zero. In this context, mixed models constitute a key methodological tool, as they allow the combination of global information with area-specific variability. However, most existing approaches rely on models with random intercepts only, implicitly assuming homogeneity in the effects of explanatory variables across domains, an assumption that is often overly restrictive in real-world applications. This thesis develops a unified methodological framework for inference in small area estimation models incorporating random slopes, thereby enabling the explicit modelling of structural heterogeneity across areas. Area-level models with discrete responses (Poisson) and continuous responses (Fay–Herriot) are studied in both cross-sectional and longitudinal settings. For each model, optimal predictors are derived, efficient parameter estimation procedures are proposed, and mean squared error estimators are developed, including both analytical and bootstrap-based approaches. The performance of the proposed methodologies is assessed through extensive simulation studies, and their practical advantages are illustrated through applications to real data, including hospital occupancy estimation in epidemic contexts and the production of disaggregated socioeconomic indicators from official surveys. Overall, this thesis extends the methodological toolkit of small area estimation by providing more flexible and realistic models for the production of reliable indicators under conditions of limited information., This PhD thesis is part of grant PRE2021-100857, funded by MCIN/AEI/10.13039/501100011033 and by the European Social Fund Plus (ESF+). This research has also been supported by the following national grants: PGC2018-096840-B-I00, PID2020-113578RB-I00, PID2023-147127OB-I00 (funded by MICIU/AEI/10.13039/501100011033 and by the European Regional Development Fund – ERDF/EU), and PID2022-136878NB-I00; by funding from the Generalitat Valenciana through Prometeo/2021/063 and CIPROM/2024/34; and by funding from the Xunta de Galicia through the Competitive Reference Groups programme (ED431C/2020/14) and the Galicia Singular Research Centre programme (ED431G/2019/01 and ED431G 2023/01), all co-financed by the ERDF. In addition, the completion of the second national research stay at Miguel Hern´andez University of Elche (UMH), Spain, was made possible thanks to the MODES research group and the CITIC. CITIC, as a centre accredited for excellence within the Galician University System and a member of the CIGUS Network, receives funding from the Department of Education, Science, Universities and Vocational Training of the Xunta de Galicia. Furthermore, it is co-financed by the European Union through the ERDF Galicia 2021–2027 Operational Programme (Ref. ED431G 2023/01). We would also like to thank the Galicia Supercomputing Center (CESGA) for providing the computational resources that enabled part of the tests and experiments conducted during this thesis., Xunta de Galicia; ED431C/2020/14, Xunta de Galicia; ED431G/2019/01, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Local Correlation Integral Approach for Anomaly Detection Using Functional Data

  • Sosa, Jorge
  • Flores Sánchez, Miguel
  • Naya, Salvador
  • Tarrío-Saavedra, Javier
[Abstract]: The present work develops a methodology for the detection of outliers in functional data, taking into account both their shape and magnitude. Specifically, the multivariate method of anomaly detection called Local Correlation Integral (LOCI) has been extended and adapted to be applied to the particular case of functional data, using the calculation of distances in Hilbert spaces. This methodology has been validated with a simulation study and its application to real data. The simulation study has taken into account scenarios with functional data or curves with different degrees of dependence, as is usual in cases of continuously monitored data versus time. The results of the simulation study show that the functional approach of the LOCI method performs well in scenarios with inter-curve dependence, especially when the outliers are due to the magnitude of the curves. These results are supported by applying the present procedure to the meteorological database of the Alternative Energy and Environment Group in Ecuador, specifically to the humidity curves, presenting better performance than other competitive methods., Xunta de Galicia; ED431C-2020-14, This work has been supported by the CITIC, that is funded by the Xunta de Galicia through the collaboration agreement between the Consellería de Cultura, Educación, Formación Profesional e Universidades and the Galician universities for the reinforcement of the research centers of the Galician University System (CIGUS). This study was also supported by the External Research Project PIE-DM-ESPOCH-2020 of the Escuela Politécnica Nacional, by the Ministerio de Ciencia e Innovación grant PID2020-113578RB-100, and the Xunta de Galicia (Grupos de Referencia Competitiva ED431C- 2020-14), all of them through the ERDF.
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Outlier Detection for Multivariate Time Series: A Functional Data Approach ®

  • López-Oriona, Ángel
  • Vilar, José
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, [Abstract] A method for detecting outlier samples in a multivariate time series dataset is proposed. It is assumed that an outlying series is characterized by having been generated from a different process than those associated with the rest of the series. Each multivariate time series is described by means of an estimator of its quantile cross-spectral density, which is treated as a multivariate functional datum. Then an outlier score is assigned to each series by using functional depths. A broad simulation study shows that the proposed approach is superior to the alternatives suggested in the literature and demonstrates that the consideration of functional data constitutes a critical step. The procedure runs in linear time with respect to both the series length and the number of series, and in quadratic time with respect to the number of dimensions. Two applications concerning financial series and ECG signals highlight the usefulness of the technique., This research has been supported by the Ministerio de Economía y Competitividad (MINECO) grants MTM2017-82724-R and PID2020-113578RB-100, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14), and the Centro de Investigación del Sistema Universitario de Galicia “CITIC” grant ED431G 2019/01; all of them through the European Regional Development Fund (ERDF). This work has received funding for open access charge by Universidade da Coruña/CISUG ., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Bootstrap-based statistical inference for linear mixed effects under misspecifications

  • Reluga, Katarzyna
  • Lombardía, María José
  • Sperlich, Stefan
[Abstract]: Linear mixed effects are considered excellent predictors of cluster-level parameters in various domains. However, previous research has demonstrated that their performance is affected by departures from model assumptions. Given the common occurrence of these departures in empirical studies, there is a need for inferential methods that are robust to misspecifications while remaining accessible and appealing to practitioners. Statistical tools have been developed for cluster-wise and simultaneous inference for mixed effects under distributional misspecifications, employing a user-friendly semiparametric random effect bootstrap. The merits and limitations of this approach are discussed in the general context of model misspecification. Theoretical analysis demonstrates the asymptotic consistency of the methods under general regularity conditions. Simulations show that the proposed intervals are robust to departures from modelling assumptions, including asymmetry and long tails in the distributions of errors and random effects, outperforming competitors in terms of empirical coverage probability. Finally, the methodology is applied to construct confidence intervals for household income across counties in the Spanish region of Galicia., The authors gratefully acknowledge support from the Swiss National Science Foundation, projects 200021-192345 and P2GEP2-195898, as well as from the Instituto Galego de Estatística who provided us with the data set. In addition, this research has been supported by MICINN grant PID2020-113578RB-I00, and by Xunta de Galicia (Grupos de Referencia Competitiva ED431C 2020/14), GAIN (Galician Innovation Agency) and the Regional Ministry of Economy, Employment and Industry grant COV20/00604 and Centro de investigación del Sistema universitario de Galicia ED431G 2019/01, all of them through ERDF. The computations were performed at the University of Geneva using Baobab and Yggdrasil HPC Service and using the computational facilities of the Advanced Computing Research Centre, University of Bristol., Xunta de Galicia; ED431C 2020/14, Xunta de Galicia; ED431G 2019/01, Switzerland. Swiss National Science Foundation; 200021-192345, Switzerland. Swiss National Science Foundation; P2GEP2-195898, Xunta de Galicia; COV20/00604
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Improving the prediction accuracy of statistical models: A new hierarchical clustering approach

  • López-Oriona, Ángel
  • Sun, Ying
  • Vilar, José
Open access publishing provided by King Abdullah University of Science and Technology (KAUST)., [Abstract]: Statisticians and machine learning practitioners frequently encounter datasets originated from multiple populations but containing the same type of measurements. In such cases, predictive analytics is typically carried out by either fitting a separate model to each dataset independently or by merging the datasets and fitting a single model to the combined data. These approaches overlook the potential existence of multiple groups of datasets associated with different underlying models, and, therefore, fail to exploit the inherent similarity between datasets to improve predictions. A third alternative is to perform pairwise comparisons between the populations before fitting the models. However, this is not always feasible, can become a very challenging task with complex models, and often does not rely on predictive accuracy. To address these issues, we propose a clustering approach designed to improve predictions in general databases. The method is based on a novel type of objective function that represents the total by-group prediction error. The clustering problem is solved using a hierarchical-type algorithm of agglomerative nature that automatically obtains the resulting clustering partition in a fully data-driven manner. An additional advantage of this procedure is that the number of clusters is treated as a variable in the minimization problem, allowing it to be determined naturally in a way that optimizes the predictive accuracy of the underlying models. Furthermore, the technique is versatile and can be used with any type of model for both regression, and classification tasks. Several simulation experiments and two real-world applications involving housing prices demonstrate that the procedure outperforms benchmark approaches in terms of predictive accuracy, Ángel López-Oriona and Ying Sun thank King Abdullah University of Science and Technology (KAUST) for its support. The research by José A. Vilar is supported by the grants PID2020-113578RB-I00 and PID2023-147127OB-I00 "ERDF/EU", funded by MCIN/AEI/10.13039/501100011033/. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2024/14) and by CITIC as a center accredited for excellence within the Galician University System and a member of the CIGUS Network, receiving subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia. Additionally, it is co-financed by the EU through the FEDER Galicia 2021-27 operational program (Ref. ED431G 2023/01)., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Subagging for bandwidth selection: a computationally efficient approach to kernel density estimation

  • Francisco-Fernández, Mario
  • Barreiro-Ures, Daniel
  • Cao, Ricardo
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG Link to the electronic supplementary material: Supplementary file 1 (pdf 277 KB) https://static-content.springer.com/esm/art%3A10.1007%2Fs00180-025-01712-4/MediaObjects/180_2025_1712_MOESM1_ESM.pdf, [Abstract]: Bandwidth selection is a central issue in kernel density estimation. For large datasets, classical selectors such as cross-validation and bootstrap become computationally intensive and may yield bandwidths with high variability. This paper proposes subagging-based versions of several popular selectors, including cross-validation, direct plug-in, and bootstrap methods. These selectors are constructed by computing bandwidths over multiple subsamples (without replacement), rescaling them, and averaging the results. We also introduce a novel regression-based approach, Regression Subbagging (RSB), which extrapolates the optimal bandwidth via a log-log regression, avoiding the need to assume a known convergence rate. We assess statistical accuracy in terms of the mean squared error of the selectors and the corresponding MISE of the resulting kernel estimators, using the optimal bandwidth as a benchmark. Computational efficiency is evaluated via parallel implementations using the parallel and foreach packages in R, reporting speedups as a function of the number of CPU cores. The results confirm that subagging improves or preserves statistical performance while yielding substantial runtime reductions, especially for demanding selectors like cross-validation and bootstrap. The RSB variant, in particular, stands out as a scalable, flexible, and robust solution. The core methods are implemented in the R package baggingbwsel, available on CRAN., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. Funding for open acess charge: Universidade da Coruña/CISUG. This research was supported by the grants PID2020-113578RB-I00 and PID2023-147127OB-I00 “ERDF/EU", funded by MCIN/AEI/10.13039/501100011033. It was also supported by the Xunta de Galicia through the program “Grupos de Referencia Competitiva" (grant ED431C-2024/14) and by CITIC, a center accredited for excellence within the Galician University System and a member of the CIGUS Network. CITIC receives subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia and is co-financed by the European Union through the FEDER Galicia 2021–27 operational program (Ref. ED431G 2023/01)., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

3D Point Cloud Semantic Segmentation Through Functional Data Analysis

  • Oviedo de la Fuente, Manuel
  • Cabo, Carlos
  • Roca-Pardiñas, Javier
  • Loudermilk, E. Louise
  • Ordóñez, Celestino
Financiado para publicación en acceso aberto: CRUE-CSIC/Springer Nature., [Abstract]: Here, we propose a method for the semantic segmentation of 3D point clouds based on functional data analysis. For each point of a training set, a number of handcrafted features representing the local geometry around it are calculated at different scales, that is, varying the spatial extension of the local analysis. Calculating the scales at small intervals allows each feature to be accurately approximated using a smooth function and, for the problem of semantic segmentation, to be tackled using functional data analysis. We also present a step-wise method to select the optimal features to include in the model based on the calculation of the distance correlation between each feature and the response variable. The algorithm showed promising results when applied to simulated data. When applied to the semantic segmentation of a point cloud of a forested plot, the results proved better than when using a standard multiscale semantic segmentation method. The comparison with two popular deep learning models showed that our proposal requires smaller training samples sizes and that it can compete with these methods in terms of prediction., This research/work was supported by MINECO, PID2020-113578RB-I00, MTM2017-82724-R and PID2020-116587GB-I00 grants, and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro de Investigación del Sistema universitario de Galicia ED431G 2019/01), all of them through the ERDF. “CITIC, a Research Center accredited by the Galician University System, is funded by the ”Consellería de Cultura, Educación e Universidade of the Xunta de Galicia. It was also funded in part by the US Department of Defense’s Strategic Environmental Research and Development Program, project no. RC19-1119. We acknowledge Tall Timbers Research Station for their support, including Dr. Robertson and Pebble Hill Plantation. The findings and conclusions in this publication are those of the authors and should not be construed to represent any official USDA or US Government determination or policy”. Carlos Cabo received funding from the UK Natural Environment Research Council (NE/T001194/1), and from the Spanish Government (Ministry of Universities) and the European Union (NextGenerationEU), within the project MU21-UP2021-030., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01, Universidad de Oviedo; MU-21-UP2021-030, United States. Department of Defense’s Strategic Environmental Research and Development Program; RC19-1119, United Kingdom. Natural Environment Research Council; NE/T001194/1
RUC. Repositorio da Universidade da Coruña
Publicación Reseña (Review)

Bootstrap prediction regions for daily curves of electricity demand and price using functional data

  • Peláez, Rebeca
  • Aneiros Pérez, Germán
  • Vilar, Juan M.
[Abstract]: The aim of this paper is to compute one-day-ahead prediction regions for daily curves of electricity demand and price. Three model-based procedures to construct general prediction regions are proposed, all of them using bootstrap algorithms. The first proposed method considers any norm for functional data to measure the distance between curves, the second one is designed to take different variabilities along the curve into account, and the third one takes advantage of the notion of depth of a functional data. The regression model with functional response on which our proposed prediction regions are based is rather general: it allows to include both endogenous and exogenous functional variables, as well as exogenous scalar variables; in addition, the effect of such variables on the response one is modelled in a parametric, nonparametric or semi-parametric way. A comparative study is carried out to analyse the performance of these prediction regions for the electricity market of mainland Spain, in year 2012. This work extends and complements the methods and results in Aneiros et al. (2016) (focused on curve prediction) and Vilar et al. (2018) (focused on prediction intervals), which use the same database as here., This research/work is part of the grants PID2020-113578RB-I00 and PID2023-147127OB-I00 “ERDF/EU”, funded by MCIN/AEI/10.13039/501100011033/. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2024/14) and by CITIC as a center accredited for excellence within the Galician University System and a member of the CIGUS Network, receives subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia. Additionally, it is co-financed by the EU through the FEDER Galicia 2021-27 operational program (ED431G 2023/01)., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Quantile Cross-Spectral Density: A Novel and Effective Tool for Clustering Multivariate Time Series

  • López-Oriona, Ángel
  • Vilar, José
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, [Abstract] Clustering of multivariate time series is a central problem in data mining with applications in many fields. Frequently, the clustering target is to identify groups of series generated by the same multivariate stochastic process. Most of the approaches to address this problem include a prior step of dimensionality reduction which may result in a loss of information or consider dissimilarity measures based on correlations and cross-correlations but ignoring the serial dependence structure. We propose a novel approach to measure dissimilarity between multivariate time series aimed at jointly capturing both cross dependence and serial dependence. Specifically, each series is characterized by a set of matrices of estimated quantile cross-spectral densities, where each matrix corresponds to a pair of quantile levels. Then the dissimilarity between every couple of series is evaluated by comparing their estimated quantile cross-spectral densities, and the pairwise dissimilarity matrix is taken as starting point to develop a partitioning around medoids algorithm. Since the quantile-based cross-spectra capture dependence in quantiles of the joint distribution, the proposed metric has a high capability to discriminate between high-level dependence structures. An extensive simulation study shows that our clustering procedure outperforms a wide range of alternative methods and exhibits robustness to noise distribution besides being computationally efficient. A real data application involving bivariate financial time series illustrates the usefulness of the proposed approach. The procedure is also applied to cluster nonstationary series from the UEA multivariate time series classification archive., This research has been supported by the Ministerio de Economía y Competitividad (MINECO) grants MTM2017-82724-R and PID2020-113578RB-100, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14), and the Centro de Investigación del Sistema Universitario de Galicia “CITIC” grant ED431G 2019/01; all of them through the European Regional Development Fund (ERDF). This work has received funding for open access charge by Universidade da Coruña/CISUG, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Statistical Procedure for Panel Block Assembly in Shipbuilding

  • Otero, Déborah
  • Cao, Ricardo
  • Blasco, Vicente
  • Brage, Álvaro
  • Tarrío-Saavedra, Javier
  • Naya, Salvador
[Abstract]: A statistical procedure to estimate the probability of successful sliding of transverse elements through the longitudinals in shipbuilding panel block assembly is proposed. It consists of developing a custom statistical solution to control the quality of shipbuilding block assembly process, which helps to meet the requirements of production time, cost, and resources consumption. This proposal addresses a critical shipbuilding challenge: the panel block assembly process, which involves inserting transverse pieces through panels containing longitudinal components. This statistical procedure estimates the probability of successful block assembly before the process starts, taking into account inputs such as panel dimensions, panel structure, and transverse stiffener. A comprehensive simulation study has been performed to evaluate the statistical procedure performance. In addition, an actual database obtained from Navantia shipyards has been used to obtain information about the mean values and dispersion of the block assembly parameters. Finally, a sensitivity analysis is applied in order to obtain information about the most critical inputs for process improvement. This statistical tool proposes an alternative to evaluate the proficiency of shipyards to perform panel block assembly process during the vessel construction. The identification of those critical variables in the panel assembly process and the quantification of their influence in the studied process are goals that have been also achieved., This research was supported by GAIN (Xunta de Galicia) and Navantia company (SEPI), in the framework of the UDC - Navantia joint Research Unit, with the project “Shipyard 4.0. The Shipyard of the Future” and Centro Mixto de Investigación (CEMI) Navantia-UDC, with reference IN853C 2022/01. This research has been also supported by the Ministerio de Ciencia e Innovación grant PID2020-113578RB-100 and PID2023-147127OB-I00, the Ministry for Digital Transformation and Civil Service under Grant TSI-100925-2023-1, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and ED431C 2024/014), and by the CITIC, also funded by the Xunta de Galicia through the collaboration agreement between the Consellería de Cultura, Educación, Formación Profesional e Universidades and the Galician universities for the reinforcement of the research centers of the Galician University System, CIGUS, with reference ED431G 2023/01. Funding for open access charge: Universidade da Coruña/CISUG., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431C 2024/014, Xunta de Galicia; ED431G 2023/01, Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG
RUC. Repositorio da Universidade da Coruña
Publicación Tesis doctoral (DoctoralThesis)

Nonparametric estimation of the probability of default in credit risk.

  • Peláez, Rebeca
[Abstract] Financial institutions are interested in knowing the probability that their clients declare themselves unable to pay the debts incurred by granting a credit. The aim of this work is to propose models to estimate this probability, called probability of default (PD), using the information provided by the credit scoring. The PD conditional on the credit scoring can be written as a transformation of the conditional survival function of the variable “time to default”. This property is used to propose new PD estimators, based on nonparametric estimators of the survival function. The time to default faces a right-censoring problem, since in the study of a set of loans, it is not possible to observe default for all of them. Consequently, censored data techniques and survival analysis are used. Given the possible existence of individuals not susceptible to default, mixture cure models are also discussed in this work. The asymptotic expression for the mean squared error and the asymptotic normality of the proposed estimators are obtained. Automatic bootstrap selectors are proposed for the smoothing parameters on which the estimators depend. The performance of the proposed techniques is analysed and compared with existing semiparametric approaches through simulation studies and illustrated by analysing bank loan data, [Resumen] A las entidades financieras les interesa conocer la probabilidad de que sus clientes se declaren incapaces de hacer frente a las deudas contraídas con la concesión de un crédito. El objetivo de este trabajo es proponer modelos para estimar esta probabilidad, denominada probabilidad de mora (PD), utilizando la información proporcionada por la puntuación crediticia. La PD condicionada a la puntuación crediticia puede escribirse como una transformaciónde la función de supervivencia condicional de la variable “tiempo hasta la mora”. Esta propiedad se utiliza para proponer nuevos estimadores de la PD, basados en estimadores no paramétricos de la función de supervivencia. El tiempo hasta el impago se enfrenta a un problema de censura por la derecha, pues en elestudio de un conjunto de créditos, no es posible observar la mora para todos ellos. Consecuentemente, se utilizan técnicas de datos censurados y análisis de supervivencia. Ante la posible existencia de individuos no susceptibles a la mora, los modelos de curación de tipo mixtura también se discuten en este trabajo. Se obtiene la expresión asintótica para el error cuadrático medio y la normalidad asintótica de los estimadores propuestos. Se proponen selectores automáticos bootstrap para los parámetros de suavizado de los que dependen los estimadores. El comportamiento de las técnicas propuestas se analiza y se compara con enfoques semiparamétricos existentes mediante estudios de simulación y se ilustra mediante el análisis de datos de préstamos bancarios., [Resumo] Ás entidades financeiras interesalles coñecer a probabilidade de que os seus clientes declárense incapaces de facer fronte ás débedas contraídas coa concesión dun crédito. O obxectivo deste traballo é propoñer modelos para estimar esta probabilidade, denominada probabilidade de morosidade (PD), utilizando a información proporcionada pola puntuación crediticia. A PD condicionada á puntuación crediticia pode escribirse como unha transformación da función de supervivencia condicional da variable “tempo ata a morosidade”. Esta propiedade utilízase para propoñer novos estimadores da PD, baseados en estimadores non paramétricos da función de supervivencia. O tempo ata a falta de pagamento enfróntase a un problema de censura pola dereita, pois no estudo dun conxunto de créditos, non é posible observar a falta de pagamento para todos eles. Consecuentemente, utilízanse técnicas de datos censurados e análise de supervivencia. Ante a posible existencia de individuos non susceptibles á morosidade, os modelos de curación de tipo mestura tamén se discuten neste traballo. Obtense a expresión asintótica para o erro cuadrático medio e a normalidade asintótica dos estimadores propostos. Propóñense selectores automáticos bootstrap para os parámetros de suavizado dos que dependen os estimadores. O comportamiento das técnicas propostas analízase e compárase con enfoques semiparamétricos existentes mediante estudos de simulación e ilústrase mediante a análise de datos de préstamos bancarios., Xunta de Galicia; ED431C-2016-015, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G/01, Xunta de Galicia; ED431G 2019/01, The PhD candidate’s research was sponsored by the Spanish Grant for Predoctoral Research Trainees RD 103/2019. The work has been partially carried out during a stay at KU Leuven (Belgium), which was supported by inMOTION Programme of grants for pre-doctoral stays Inditex-UDC 2021. This research has also been partially supported by MINECO Grants MTM2017-82724-R and PID2020-113578RB-100, and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2016-015 and ED431C-2020-14 and Centro Singularde Investigación de Galicia ED431G/01 and ED431G 2019/01), all of them through the ERDF.
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

A Nonparametric Bootstrap Method for Heteroscedastic Functional Data

  • Fernández-Casal, Rubén
  • Castillo-Páez, Sergio
  • Flores Sánchez, Miguel
Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature., The pre-processed data are supplied with the R package npfda, as the ozone data set. The original data can be downloaded from https://uk-air.defra.gov.uk/data. The code used to apply the proposed methodology to the pollution data and the results generated are included in the supplementary material., [Absctract]: The objective is to provide a nonparametric bootstrap method for functional data that consists of independent realizations of a continuous one-dimensional process. The process is assumed to be nonstationary, with a functional mean and a functional variance, and dependent. The resampling method is based on nonparametric estimates of the model components. Numerical studies were conducted to check the performance of the proposed procedure, by approximating the bias and the standard error of two estimators. A practical application of the proposed approach to pollution data has also been included. Specifically, it is employed to make inference about the annual trend of ground-level ozone concentration at Yarner Wood monitoring station in the United Kingdom. Supplementary material to this paper is provided online., This work has been supported by grant PID2020-113578RB-I00, funded by MCIN/AEI/10.13039/501100011033. The research of Rubén Fernández-Casal has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020/14) and by CITIC that is supported by Xunta de Galicia, convenio de colaboración entre la Consellería de Cultura, Educación, Formación Profesional e Universidades y las universidades gallegas para el refuerzo de los centros de investigación del Sistema Universitario de Galicia (CIGUS). The research of Sergio Castillo Páez has also been supported by the Universidad de las Fuerzas Armadas ESPE, from Ecuador., Xunta de Galicia; ED431C-2020-14
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Nonparametric Conditional Risk Mapping Under Heteroscedasticity

  • Fernández-Casal, Rubén
  • Castillo-Páez, Sergio
  • Francisco-Fernández, Mario
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature, [Absctract]: A nonparametric procedure to estimate the conditional probability that a nonstationary geostatistical process exceeds a certain threshold value is proposed. The method consists of a bootstrap algorithm that combines conditional simulation techniques with nonparametric estimations of the trend and the variability. The nonparametric local linear estimator, considering a bandwidth matrix selected by a method that takes the spatial dependence into account, is used to estimate the trend. The variability is modeled estimating the conditional variance and the variogram from corrected residuals to avoid the biasses. The proposed method allows to obtain estimates of the conditional exceedance risk in non-observed spatial locations. The performance of the approach is analyzed by simulation and illustrated with the application to a real data set of precipitations in the USA.Supplementary materials accompanying this paper appear on-line., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. Research of A. Meilán-Vila and M. Francisco-Fernández has been supported by MINECO (Grant MTM2017-82724-R), MICINN (Grant PID2020-113578RB-I00), and by Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro de Investigación del Sistema Universitario de Galicia ED431G 2019/01), all of them through the ERDF. Research of R. M. Crujeiras has been supported by MICINN (Grant PID2020-116587GB-I00), and by Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2021-24), all of them through the ERDF., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01, Xunta de Galicia; ED431C-2021-24
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Predicting lifetime of adhesive bonds for naval steel by time-temperature superposition

  • Rodríguez-Dopico, Francisco J.
  • Álvarez García, Ana
  • Tarrío-Saavedra, Javier
  • Meneses Freire, Antonio
  • Naya, Salvador
[Abstract] There is a lack of knowledge about the long-term behaviour of adhesive joints in the marine environment, and for hence, its reliability. The life cycle expected for a ship is twenty-five years and conduct duration tests would last decades. The importance and originality of this study are that it provides a methodology for predicting the durability of adhesive bonds. For this purpose, three adhesives for bonding naval steel were tested at several temperatures for determination of their shear strength on standard single-lap-joint specimens. Subsequently, using the time temperature superposition principle, these results were combined into one master curve for each adhesive that can be considered as a prediction method of the durability in a long-term period. The usual parametric models of Arrhenius and Williams-Landel-Ferry were used to obtain the master curve for each adhesive by manual shifting. The results were compared with those obtained applying an open-source software developed by the authors in R language, which specifically implements a non-parametric methodology based on the shift of the first derivative curves, following parameters of Explainable Artificial Intelligence. It was found a good correlation between both methodologies, supporting the durability lifetime obtained and therefore the applicability on ships of this adhesive technology., The authors wish to acknowledge and thank the support provided by Krafft S.L.U. Spain (ITW Performance Polymers), Sarsch Adhesives, S.L. Spain (Engineered Bonding Solutions, LLC, Florida, United States) and Masterbond Inc. United States. The research of Javier Tarrío Saavedra and Salvador Naya has been supported by the Ministerio de Ciencia e Innovación grant PID2020-113578RB-100 and PID2023-147127OB-I00, the Ministry for Digital Transformation and Civil Service under Grant TSI-100925-2023-1, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and ED431C 2024/014), and by the CITIC, also funded by the Xunta de Galicia through the collaboration agreement between the Consellería de Cultura, Educación, Formación Profesional e Universidades and the Galician universities for the reinforcement of the research centers of the Galician University System, CIGUS, with reference ED431G 2023/01., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431C 2024/014, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Nonparametric estimation for a functional-circular regression model

  • Meilán-Vila, Andrea
  • Crujeiras-Casais, Rosa M.
  • Francisco-Fernández, Mario
Financiado para publicación en acceso aberto: CRUE-CSIC, [Abstract]: Changes on temperature patterns, on a local scale, are perceived by individuals as the most direct indicators of global warming and climate change. As a specific example, for an Atlantic climate location, spring and fall seasons should present a mild transition between winter and summer, and summer and winter, respectively. By observing daily temperature curves along time, being each curve attached to a certain calendar day, a regression model for these variables (temperature curve as covariate and calendar day as response) would be useful for modeling their relation for a certain period. In addition, temperature changes could be assessed by prediction and observation comparisons in the long run. Such a model is presented and studied in this work, considering a nonparametric Nadaraya–Watson-type estimator for functional covariate and circular response. The asymptotic bias and variance of this estimator, as well as its asymptotic distribution are derived. Its finite sample performance is evaluated in a simulation study and the proposal is applied to investigate a real-data set concerning temperature curves., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. Research of A. Meilán-Vila and M. Francisco-Fernández has been supported by MINECO (Grant MTM2017-82724-R), MICINN (Grant PID2020-113578RB-I00), and by Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro de Investigación del Sistema Universitario de Galicia ED431G 2019/01), all of them through the ERDF. Research of R. M. Crujeiras has been supported by MICINN (Grant PID2020-116587GB-I00), and by Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2021-24), all of them through the ERDF., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01, Xunta de Galicia; ED431C-2021-24
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Characterization of Epoxy Adhesive for Marine Applications

  • Rodríguez-Dopico, Francisco J.
  • Álvarez García, Ana
  • Tarrío-Saavedra, Javier
[Abstract] The shipping industry faces the challenge of expanding its capabilities and mitigating its environmental impact. The utilization of adhesive bonding techniques has the potential to facilitate the construction of lighter ships with a reduced carbon footprint. In this work, a complete methodology has been established to study the properties of an adhesive bond between naval steel and the behavior of the bulk adhesive when introduced into seawater. Thermogravimetric analysis (TGA) has been used to determine the thermal properties. Single-lap-joint and tensile tests have been performed showing the negative influence of temperature on the mechanical properties. Seawater absorption has been studied by gravimetric tests. Differential scanning calorimetry (DSC), Fourier transform infrared spectroscopy (FTIR), and mechanical bulk tests have been carried out before and after aging. The water absorption is Fickian, and when a constant load is applied, the absorption occurs faster and in greater quantity, decreasing the stiffness and strength. An increase in the glass transition temperature was observed in the absence of a load, with no discernible change upon the application of load condition. Also, aged specimens with constant load exhibited a reduction in water content. These phenomena may be attributed to several diffusion processes., This research was made possible by the support of the Campus Industrial de Ferrol (CIF) and the Centro de Investigación en Tecnologías Navales e Industriales (CITENI) of the University of A Coruña. The authors wish to acknowledge and thank Masterbond Inc., United States, for their support. The research of Javier Tarrío Saavedra has been supported by the Ministerio de Ciencia e Innovación grant PID2020-113578RB-100 and PID2023-147127OB-I00, the Ministry for Digital Transformation and Civil Service under Grant TSI-100925-2023-1, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and ED431C 2024/014), and by the CITIC, also funded by the Xunta de Galicia through the collaboration agreement between the Consellería de Cultura, Educación, Formación Profesional e Universidades and the Galician universities for the reinforcement of the research centers of the Galician University System, CIGUS, with reference ED431G 2023/01., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431C 2024/014, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Cost-sensitive thresholding over a two-dimensional decision region for fraud detection

  • C-Rella, Jorge
  • Cao, Ricardo
  • Vilar, Juan M.
[Absctract]: Credit fraud poses a challenging task in terms of detection. It can result in significant losses depending on the amount, so a cost-sensitive perspective needs to be taken. Classical approaches focus on estimating the probability of fraud and selecting a decision threshold, but they often fail to consider the transaction amount or account for the cumulative losses incurred within the sample. Consequently, these approaches can result in sub-optimal strategies. A new thresholding approach is proposed, based on the construction of a two-dimensional decision space with an estimated probability and the credit amount. This expansion allows more freedom for the optimal classification rule search, which is performed with a new algorithm. The proposed method generalizes previous approaches, so an improvement is consistently achieved. In addition, it allows a restricted search. This is shown in a study of two real data sets, comparing the results obtained by a wide range of classifiers., This research has been supported by MICINN Grant PID2020-113578RB-I00 and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro de Investigación del Sistema Universitario de Galicia ED431G 2019/01), all of them through the European Regional Development Fund (ERDF). The first author was financed by the Axencia Galega de Innovación Grant 14-IN606D-2021-2607768., Xunta de Galicia; ED431C-2020/14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Instance-dependent cost-sensitive parametric learning

  • C-Rella, Jorge
  • Claeskens, Gerda
  • Cao, Ricardo
  • Vilar, Juan M.
[Abstract]: Instance-dependent cost-sensitive learning addresses classification problems where each observation has a different misclassification cost. In this paper, we propose cost-sensitive parametric models to minimize the expectation of losses. A loss function incorporating the misclassification costs is defined, which serves as the objective function for obtaining cost-sensitive parameter estimators. The consistency and asymptotic normality of these estimators are established under general conditions, theoretically demonstrating their good performance. Additionally, we derive bounds for the bias introduced when regularizing the optimization problem, which is generally necessary in practice. To conclude, the effectiveness of the proposed estimators is evaluated through an extensive novel simulation study and the analysis of five real data sets, further demonstrating their proficiency in practical settings., This research has been financed by the Grant PID2020-113578RB-I00 and PID2023-147127OB-I00 ”ERDF/EU”, funded by the sponsor MCIN/AEI/10.13039/501100011033/, Spain. It has also been supported by the Xunta de Galicia, Spain (Grupos de Referencia Competitiva ED431C-2024/14) and by CITIC as a center accredited for excellence within the Galician University System and a member of the CIGUS Network, receives subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia. Additionally, it is co-financed by the EU through the FEDER Galicia 2021-27 operational program (Ref. ED431G 2023/01). The first author was financed by the Axencia Galega de Innovación Industrial PhD Grant, Spain 14-IN606D-2021-2607768 and the INDITEX-UDC mobility grant, Spain 04.00.47.00.01 422D 48001., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01, Xunta de Galicia; IN606D-2021-2607768
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Optimization of Functional Diagnostic Test: The Effect of Kernel Method as an Estimator of ROC Curve

  • Estévez-Pérez, G.
This is an Accepted Manuscript version of the article, accepted for publication in Journal of Statistical Computation and Simulation., [Abstract] Technical development over the last few decades has resulted in the emergence of complex data, in many cases functional data (FD). This type of data can emerge in many medical studies which are geared towards detecting diseases, predicting their course or evaluating the response to a therapy, to name a few. Thus, it is very useful to have statistical methods enabling us to evaluate diagnostic tests based on functional biomarkers. In fact, a diagnostic test that uses functional variables as biomarkers has been developed recently. Their authors proposed a functional version of ROC analysis, resulting in an empirical estimate of the functional ROC curve. In order to improve this methodology, the present paper proposes a procedure to obtain a smooth version of non-parametric estimator of the ROC curve. In addition, a comprehensive simulation study lets to investigate the discriminatory and predictive abilities of the resulting functional diagnostic test. Two examples with real medical data illustrate the approach developed: one deals with gene expression levels for tumoural/normal samples of prostate cancer; the other dataset is about white matter structures in the brain in multiple sclerosis patients., This research was supported by MICINN, Spain grant PID2020-113578RB-I00 and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro de Investigación del Sistema Universitario de Galicia ED431G 2019/01), all of them through the ERDF, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Prediction of Poverty Proportions Under Random Regression Coefficients Two-Fold Fay-Herriot Models

  • Diz-Rosales, Naomi
  • Lombardía, María José
  • Morales, Domingo
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG Downloadable supplementary data., [Abstract] Using a two-fold Fay-Herriot model with random intercepts and random regression coefficients, we develop area-level predictors for poverty proportions and propose both analytic and bootstrap-based estimators for the mean squared error. Model parameters are estimated via residual maximum likelihood, and empirical best linear unbiased predictors are used for random effects. To assess the performance of the estimation algorithm, predictors, and mean squared error estimators, we conduct simulation studies. The proposed methodology is applied to data from the 2022 Spanish Living Conditions Survey to estimate poverty proportions at the provincial level, disaggregated by gender and age group. This study offers a robust and innovative framework for small-area estimation, enabling detailed poverty mapping with accurate estimates and reliable uncertainty measures., This work was supported by the Ministry of Science and Innovation and the State Research Agency of the Spanish Government through the European Regional Development Fund (PID2020-113578RB-I00, funded by MCIN/AEI/10.13039/501100011033/, PID2023-147127OB-I00 funded by MICIU/AEI/10.13039/501100011033/and “FEDER/UE,” PID2022-136878NB-I00 and PRE2021-100857 to Naomi Diz-Rosales, funded by MCIN/AEI/10.13039/501100011033/and ESF+, as part of the Spanish Grant for Predoctoral Research Trainees RD 103/2019); by the Conselleria d’Innovació, Universitats, Ciéncia i Societat Digital of the Generalitat Valenciana (Prometeo/2021/063 and Prometeo CIPROM2024_34); by the Consellería de Cultura, Educación, Formación Profesional e Universidades of the Xunta de Galicia through the European Regional Development Fund (Competitive Reference Groups ED431C-2024/14); and by Centro de Investigación en Tecnologías de la Información y las Comunicaciones (CITIC), as a center accredited for excellence within the Galician University System and a member of the CIGUS Network, receives subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia. Additionally, it is co-financed by the EU through the FEDER Galicia 2021–27 operational program (Ref. ED431G 2023/01). We thank the Centro de Supercomputación de Galicia (CESGA) for providing their services for part of the simulations in this work. Funding for open access charge: Universidade da Coruña/CISUG., Generalitat Valenciana; Prometeo/2021/063, Generalitat Valenciana; Prometeo CIPROM2024_34, Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01, https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/jssam/14/4/10.1093_jssam_smaf032/1/smaf032_supplementary_data.zip?Expires=1794983846&Signature=LxM~ExufH0Hqw1vJpckRk9l30ZULR-Wu-j-YX0f7QVSQwo1AocOiNpJ0~CtifYIcs78vnz963Yf4EHc0l5d5Lzga3lItqCQrlHox6K66CK4mKgOFJKckx0bvuzH7isPI7e--kGVOXJYudAfP3ryckThWiitCrPWUGfxVF-L-qB22ZTGd2-kXC4ZQqXIR3GCp6C3G6TWMHySOFEOIoNaSE0GRU2hGcyw4P2p6INa~jKNbNx9PvNqqXxPLmFZJXSj6kvIIf5HJzDjkQL8yDU9E9BB8LaP9pYNWsxwWpjCODtG2hdn7AMx9aVqJOkdTfYUnhaqFgwueV2xjKBjjbNCA8w__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

A spatial statistical approach to estimate bus stop demand using GIS-processed data

  • Montero-Lamas, Yaiza
  • Fernández-Casal, Rubén
  • Varela-García, Francisco-Alberto
  • Orro, Alfonso
  • Novales, Margarita
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, [Abstract:] This study integrates the fields of geography, urban transit planning, and statistical learning to develop a sophisticated methodology for predicting bus demand at the stop level. It uses a Generalized Additive Model that captures non-linear relationships and incorporates spatial dependence, improving traditional methods. It showcases a high predictive capacity with a pseudo R-squared of 0.79 during its validation, ensuring substantial explanatory power for new observations. A large number of variables, including land-use characteristics, socioeconomic factors, and transit supply, are analysed. These widely available predictors facilitate the transferability of the methodology to other urban areas. Transit supply predictor considers the number of annual trips per stop and area as well as the location of stops along the lines that serve them. GIS processing of the data allows the calculation of variables within the areas of influence of each stop, obtained by following the walkable street network. For the case study, the presence of universities, hospitals, and lodgings areas, as well as inhabitants and ratio of bus trips show a positive impact on bus demand. This geo-analysis process employs accurate disaggregated data, such as information on uses in each building, as well as methods for assigning socioeconomic information from local areas to residential buildings. This study highlights the complex relationship between the location of transit network stops, both along the bus line and in terms of geographical proximity, their transit supply, and its surrounding factors. The results indicate that there is spatial dependence for stops less than 1.15 km apart. The developed methodology provides reliable information to transit network planners for decision making. Specifically, this proposed methodology can contribute to designing new routes, optimizing stop locations, and estimating the impact of changes in the transit network or urban planning on bus demand. All these improvement measures promote sustainable urban mobility, consequently fostering environmental and social benefits., The authors would like to thank Compañía de Tranvías de La Coruña and Concello da Coruña for providing the data required to prepare this paper. This work was funded by grants PID2021-128255OB-I00 and PRE2019-089651, funded by MCIN/AEI/10.13039/501100011033 and by ERDF/EU and ESF/EU. The research of Rubén Fernández-Casal has also been supported by Grant PID2020-113578RB-I00, funded by MCIN/AEI/10.13039/501100011033, by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020/14) and by CITIC that is supported by Xunta de Galicia, convenio de colaboración entre la Consellería de Cultura, Educación, Formación Profesional e Universidades y las universidades gallegas para el refuerzo de los centros de investigación del Sistema Universitario de Galicia (CIGUS). Funding for open access charge: Universidade da Coruña/CISUG., Xunta de Galicia; ED431C-2020/14
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Combined effect of seawater and load on methacrylate adhesive

  • Rodríguez-Dopico, Francisco J.
  • Carbas, R.J.C.
  • Borges, C.S.P.
  • Tarrío-Saavedra, Javier
  • Da Silva, L.F.M.
  • Álvarez García, Ana
[Abstract]: Although the shipbuilding industry is constantly demanding new advanced joining solutions, adhesive technology is not as developed in the marine as compared to other industries. The main reason is the lack of specific knowledge that guarantees the durability of the bonded joints in optimal conditions during the life cycle of a ship. This work simulates in the laboratory a marinelike environment by immersing an adhesive in seawater and subjecting it to constant loading. The objective is to characterize the seawater absorption behavior and its consequences on the mechanical, thermal, and chemical properties of the adhesive after this aging process. Seawater ingress was determined through gravimetric tests at several load conditions of the tensile strength of the adhesive. Besides, absorption process was studied using Fick’s Law, determining the diffusion coefficients. The thermal behavior was monitored with differential scanning calorimetry (DSC) and the chemical degradation was analyzed using Fourier transform infrared spectroscopy (FTIR). Also, the mechanical properties were determined by tensile tests. The surface of the adhesive (dried) was studied by Scanning Electron Microscopy (SEM) technique and the porosity was measured by physisorption with a high-performance adsorption analyzer. A numerical simulation was developed using Darcy’s Law combined with continuity equation. The results show that application of loads and immersion in seawater until full saturation of seawater improve the mechanical properties of the adhesive, but it affects negatively to the glass transition temperature. This should be considered when designing adhesive bonding joints on ships., Xunta de Galicia; ED431C-2020-14, The research of Javier Tarrío Saavedra has been supported by the Ministerio de Ciencia e Innovación grant PID2020-113578RB- 100, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14), and by the CITIC, also funded by the Xunta de Galicia through the collaboration agreement between the Consellería de Cultura, Educaci´on, Formaci´on Profesional e Universidades and the Galician universities for the reinforcement of the research centers of the Galician University System (CIGUS).
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Flexible multi-class cost-sensitive thresholding

  • C-Rella, Jorge
  • Vilar, Juan M.
[Abstract]: Classification involves categorizing input data into predefined classes based on their characteristics. Thresholding methods predict the optimal class for an observation given a score and a missclassification error cost specification. In multi-class classification, existing algorithms assume that a score is available for each possible response. However, there are scenarios where more classes can be predicted than the underlying response variable has. This paper extends the flexibility of the 2-DDR algorithm introduced by C-Rella et al. (Inf Sci 657:119956;2024) to the multi-class classification problem. The proposed method predicts the optimal classification in cost-sensitive multi-class problems considering a single score fitted over a binary variable, a problem not previously studied. Furthermore, a more efficient version of the algorithm is proposed. The good performance of the proposed multi-class method is demonstrated through extensive simulations and the analysis of four real data sets., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This research is part of the grants PID2020-113578RB-I00 and PID2023-147127OB-I00 "ERDF/EU", funded by MCIN/AEI/10.13039/501100011033/. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2024/14) and by CITIC as a center accredited for excellence within the Galician University System and a member of the CIGUS Network, receives subsidies from the Department of Education, Science, Universities, and Vocational Training of the Xunta de Galicia. Additionally, it is co-financed by the EU through the FEDER Galicia 2021-27 operational program (Ref.ED431G 2023/01). The first author was financed by the Axencia Galega de Innovación Grant 14-IN606D-2021-2607768., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01, Xunta de Galicia; 14-IN606D-2021-2607768
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Validity and reliability of an offline ultrasound measurement of bladder base displacement in women

  • Martínez-Bustelo, Sandra
  • Ferri Morales, Asunción
  • Castillo-García, Fernando J.
  • Madrid, Antonio
  • Jácome, M. A.
[Abstract] The effect of different exercises on the position of pelvic organs in women has not been sufficiently assessed. The objective was to analyze the validity and reliability of a new two-dimensional ultrasound algorithm to measure offline the displacement of the bladder base during abdominal exercises. This algorithm could be a useful method to future studies in determine the most appropriate exercises in sports and in rehabilitative program for the pelvic floor in women. All subjects were tested by transverse transabdominal ultrasound. The measurements were conducted offline using a customized code written in MATLAB (Ecolab) for image-processing, and manually on the ultrasound monitor using electronic calipers. The agreement was assessed with a paired t-test, Pearson’s linear correlation coefficient (r), the Lin’s concordance correlation coefficient (CCC), the intraclass correlation coefficient ICC (A,2) and a Bland–Altman plot. The reliability was confirmed by the interdays intra-rater ICC coefficient. The results were that Ecolab and ultrasound transducer measures did not differ statistically (p = 0.246). Furthermore, both methods showed a very strong relationship, and the Ecolab demonstrated to be a valid and reliable method. We concluded that Ecolab seemed to be a valid and reliable tool to assess the effect of abdominal contractions in the female pelvic floor., This research was funded by COFIGA (Colegio Oficial de Fisioterapeutas de Galicia, Research Grant 2018–2019 modality B. esides, I was supported by MICINN grant PID2020-113578RB-I00, and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020/14 and Centro de Investigación del Sistema Universitario de Galicia ED431G 2019/01), both of them through the ERDF, Xunta de Galicia; ED431C-2020/14, Xunta de Galicia; ED431G 2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Predicting rainfall and irrigation requirements of corn in Ecuador

  • Flores Sánchez, Miguel
  • Llambo, Ángel
  • Loza, Danilo
  • Naya, Salvador
  • Tarrío-Saavedra, Javier
[Abstract]: This work is a case study whose objective is prediction of irrigation needs of corn crops in different regions of Ecuador; being this a fundamental basic food for the country's economy, as in the remaining countries of the Andean area. The proposed methodology seeks to help improving the quality of corn crop. Specifically, we propose the application of regression models, within the framework of Functional Data Analysis (FDA), to predict the amount of rainfall (scalar response variable) in the places with the highest production of corn in Ecuador, as a function of functional covariates such as temperature and wind speed. From the estimation of the amount of rainfall, effective precipitation is calculated. This is the fraction of water used by the crops, from which the value of real evapotranspiration or ETc is obtained and, more importantly, the irrigation requirements at each stage of the corn crop, for its adequate physiological development. Application of regression models based on functional basis, Functional Principal Components (FPC) or Functional Partial Least Squares (FPLS) for scalar response variable, allows us to use the information of variables such as wind speed and temperature (of functional nature) in a better way than using multivariate models, for predicting the amount of rainfall, obtaining, as a result, very explicative models, defined by a high goodness of fit (R2=0.97, with 6 significant parameters and an error of 0.14) and practical utility. The model has been also applied to North Peru regions, obtaining rainfall prediction errors between 9% and 22%. Thus, the geographical limitations of the model could be the Andean regions with similar climate. In addition, this study proposes the application of FDA exploratory analysis and FDA outlier detection techniques as a common and useful practice in the specific domain of rainfall prediction studies, prior to applying the regression models., This research has been supported by Ministerio de Ciencia e Innovación grant PID 2020-113578RB-100 and by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14 and Centro de Investigación del Sistema universitario de Galicia ED431G2019/01), all of them through the ERDF. The authors would like to thank its support to the Escuela Politécnica Nacional, specifically to “Entidad Operativa Desconcentrada Unidad de Gestión de la Investigación y Proyección Social-EOD-UGIPS” belonging to “VICERRECTORADO DE INVESTIGACIÓN, INNOVACIÓN Y VINCULACIÓN DIRECCIÓN DE INVESTIGACIÓN”. Moreover, we would like to strongly thank all the guidance and comments of the two reviewers and the editor. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G2019/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Considerations on the programmed functional life (one generation) of a green artificial reef in terms of the sustainability of the modified ecosystem

  • Carral Couce, Luis
  • Tarrío-Saavedra, Javier
  • Cartelle Barros, Juan José
  • Camba, C.
  • Ramil, Alberto
  • Álvarez-Feal, José-Carlos
[Abstract] The installation of artificial reefs serves to enhance marine ecosystems, although it also modifies them. These changes do not have to be irreversible, since it is possible to treat the functional life of an artificial reef (AR) as a variable factor to be determined, with the objective of contributing to the sustainability of the ecosystem. The quest for sustainability does not end with the manufacture and installation of the AR units. It is also necessary to analyse the sustainability of the modified ecosystem, through the production of services. This leads to consider the medium-term return of the ecosystem to its initial state, once the functional life of the ARs expires. This paper presents and justifies an AR design/composition for limited functional life. It is the result of acting on the base material, the concrete, with the objective of limiting the useful life to one social generation. Four different dosages were proposed for such a purpose. They were subjected to mechanical tests (compressive strength and absorption after immersion), including an innovative abrasion-resistant one. The results allow estimating the functional life of the four types of concrete from the design variables (density, compactness, and quantity of water and cement as well as its relation). To this end linear regression models and clustering techniques were applied. The described procedure leads to an AR design for limited functional life., The authors gratefully acknowledge the financial support from the regional government of Galicia, Xunta de Galicia, through the project CN-10MMA003CT. This study was also funded through the collaboration agreement between Xunta de Galicia, Universidade da Coruña and the Universidade da Coruña Foundation (FUAC) to give continuity to the previous project. The research of Javier Tarrío Saavedra has been supported by the Ministerio de Ciencia e Innovación grant PID2020-113578RB-100, the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14), all of them through the ERDF, and by the CITIC, that is funded by the Xunta de Galicia through the collaboration agreement between the Consellería de Cultura, Educación, Formación Profesional e Universidades and the Galician universities for the reinforcement of the research centers of the Galician University System (CIGUS). The authors would also like to thank the Editor and the anonymous reviewers for their interesting comments., Xunta de Galicia; CN-10MMA003CT, Xunta de Galicia; ED431C-2020-14
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Nonparametric Estimation in Mixture Cure Models with Covariates

  • López-Cheda, Ana
  • Peng, Yingwei
  • Jácome, M. A.
This version of the article has been accepted for publication, after peer review (when applicable) and is subject to Springer Nature’s AM terms of use, but is not the Version of Record and does not reflect post-acceptance improvements, or any corrections. The Version of Record is available online at: https://doi.org/10.1007/s11749-022-00840-z, [Abstract] Nonparametric estimation methods for the cure rate and the distribution of the failure time of uncured subjects with covariates for censored survival data have attracted much attention in the last few years. To model the effects of covariates on the distribution of the failure time of uncured subjects, existing works assume that the cure rate is a constant or depends on the same covariate as the distribution of uncured subjects. In this paper, we review the nonparametric estimation methods in the context of the mixture cure model and propose a new nonparametric estimator for the distribution of uncured subjects that relaxes the assumption used in the existing works. The estimation is based on the EM algorithm, which is readily available for mixture cure models, and is strongly consistent. The finite sample performance of the proposed estimator is assessed and compared with existing methods in a simulation study. Finally, the nonparametric estimation methods are employed to model the effects of some covariates on the time to bankruptcy among commercial banks insured by the Federal Deposit Insurance Corporation during the first quarter of 2006., ALC was sponsored by the BEATRIZ GALINDO JUNIOR Spanish grant from Ministerio de Ciencia, Innovación y Universidades with reference BGP18/00154. ALC and MAJ acknowledge partial support by the MINECO Grant MTM2017-82724-R (EU ERDF support included), the MICINN Grant PID2020-113578RB-I00, and partial support of Xunta de Galicia (Centro Singular de Investigación de Galicia accreditation ED431G 2019/01 and Grupos de Referencia Competitiva ED431C-2020-14 and ED431C2016-015) and the European Union (European Regional Development Fund - ERDF). YP’s work was partially supported by a Discovery grant from the Natural Sciences and Engineering Research Council of Canada. The authors thank Alessandro Beretta and Cédric Heuchenne for their assistance in obtaining the bank data analyzed in Sect. 7, Xunta de Galicia; ED431G 2019/01, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431C2016-015
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

A cardiotoxicity dataset for breast cancer patients

  • Piñeiro Lamas, Beatriz
  • López-Cheda, Ana
  • Cao, Ricardo
  • Ramos-Alonso, Laura
  • González-Barbeito, Gabriel
  • Barbeito Caamaño, Cayetana
  • Bouzas-Mosquera, Alberto
[Abstract] This dataset is a result of the collaboration between the University of A Coruña and the University Hospital of A Coruña. It contains information about 531 women diagnosed with HER2+ breast cancer, treated with potentially cardiotoxic oncologic therapies. These treatments can cause cardiovascular adverse events, including cardiac systolic dysfunction, the development of which has important clinical and prognostic implications. The availability of good predictors may enable early detection of these cardiac problems. Variables such as age, weight and height are available for each patient, as well as some measures obtained from echocardiography, a technique used prior and during the treatment to check the structure and function of the heart. Among them, there is a functional variable that measures the myocardial velocity during the cardiac cycle. For patients that experienced cancer therapy-related cardiac dysfunction during the treatment period, time until its appearance is known. This dataset aims to enable the scientific community in conducting new research on this cardiovascular side effect., B.P.L. was sponsored by Axudas Predoutorais da Xunta de Galicia, with reference ED481A-2020/290. A.L.C. was sponsored by the BEATRIZ GALINDO JUNIOR Spanish Grant from MICINN (Ministerio de Ciencia e Innovación) with code BGP18/00154. B.P.L., A.L.C. and R.C. acknowledge partial support by the MICINN Grant PID2020-113578RB-I00 and partial support of Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14). B.P.L., A.L.C. and R.C. acknowledge the support received from the Centro de Investigación de Galicia “CITIC”, funded by Xunta de Galicia and the European Union European Regional Development Fund (ERDF)-Galicia 2014–2020 Program, Grant ED431G 2019/01., Xunta de Galicia; ED481A-2020/290, Xunta de Galicia; ED431C-2020-14, Xunta de Galicia; ED431G 2019/01
DIGITAL.CSIC. Repositorio Institucional del CSIC
Publicación Artículo científico (JournalArticle)

Influence of Storage Time on the DNA Integrity and Viability of Spermatozoa of the Spider Crab Maja brachydactyla

  • Rodríguez-Pena, Elba
  • Suárez, Diego
  • Estévez, Maria Graciela
  • Verísimo-Amor, Patricia
  • Barreira, Noelia
  • Fernández, Luis
  • González-Tizón, Ana
  • Martínez-Lage, Andrés
Natural populations of the spider crab Maja brachydactyla constitute a fishery resource of great economic importance in many countries. As in the rest of eubrachyurans, the females of this species have ventral-type seminal receptacles where they store sperm from copulations. Sperm can be stored in these structures for months and even years before egg fertilisation, with the consequent degradation of the sperm cells during the time. In this work, we analyse the viability and the possible genetic damage in sperm accumulated in the seminal receptacles of M. brachydactyla females as a function of the storage time (from 0 to 14 months) using the comet assay technique. On one hand, we developed an algorithm for comet image analysis that improves the comet segmentation compared with the free software Open comet v1.3.1 (97% vs. 76% of detection). In addition, our software allows the manual modification of the contours wrongly delimited via the automatic tool. On the other hand, our data show a sharp decline in sperm viability and DNA integrity in the first four months of storage, which could lead to a decrease in the fecundity rate and/or viability of the embryos or larvae from the second and third clutches of the annual cycle if the repair capacity in these gametic cells is low., This research was funded by the Ministerio de Ciencia, Innovación y Universidades, grant number PID2019-109194RB-I00 MAR; and by the Ministerio de Ciencia e Innovación, grant number PID2020-113578RB-I00. It has also been supported by the Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020/14)., Peer reviewed
Academica-e. Repositorio Institucional de la Universidad Pública de Navarra
Publicación Artículo científico (JournalArticle)

Estimation of patient flow in hospitals using up-to-date data: application to bed demand prediction during pandemic waves

  • García de Vicuña Bilbao, Daniel
  • López-Cheda, Ana
  • Jácome, María Amalia
  • Mallor Giménez, Fermín
Hospital bed demand forecast is a first-order concern for public health action to avoid healthcare systems to be overwhelmed. Predictions are usually performed by estimating patients flow, that is, lengths of stay and branching probabilities. In most approaches in the literature, estimations rely on not updated published information or historical data. This may lead to unreliable estimates and biased forecasts during new or non-stationary situations. In this paper, we introduce a flexible adaptive procedure using only near-real-time information. Such method requires handling censored information from patients still in hospital. This approach allows the efficient estimation of the distributions of lengths of stay and probabilities used to represent the patient pathways. This is very relevant at the first stages of a pandemic, when there is much uncertainty and too few patients have completely observed pathways. Furthermore, the performance of the proposed method is assessed in an extensive simulation study in which the patient flow in a hospital during a pandemic wave is modelled. We further discuss the advantages and limitations of the method, as well as potential extensions., DGV and FM acknowledge the support by grant PID2020-114031RB-I00 (AEI, FEDER EU) and by the Government of Navarre, 0011-3597-2020-000003 (COVID). ALC was sponsored by the BEATRIZ GALINDO JUNIOR Spanish Grant from MICINN (Ministerio de Ciencia e Innovación) with code BGP18/00154. ALC and MAJ acknowledge partial support by the MICINN Grant PID2020-113578RB-I00 and partial support of Xunta de Galicia (Grupos de Referencia Competitiva ED431C-2020-14). ALC and MJ wish to acknowledge the support received from the Centro de Investigación de Galicia "CITIC", funded by Xunta de Galicia and the European Union European Regional Development Fund (ERDF)-Galicia 2014-2020 Program, by grant ED431G 2019/01.