Pasar al contenido principal

INFERENCIA ESTADISTICA UTILIZANDO METODOS FLEXIBLES PARA DATOS COMPLEJOS: TEORIA Y APPLICACIONES

PID2023-147127OB-I00

•
Nombre agencia financiadora Agencia Estatal de Investigación
Acrónimo agencia financiadora AEI
Programa Programa Estatal para Impulsar la Investigación Científico-Técnica y su Transferencia
Subprograma Subprograma Estatal de Generación de Conocimiento
Convocatoria Proyectos de I+D+i (Generación de Conocimiento y Retos Investigación)
Año convocatoria 2023
Plan Estatal 2021-2023 Plan Estatal de Investigación Científica y Técnica y de Innovación
Centro beneficiario UNIVERSIDADE DA CORUÑA

Resultados relacionados

Resultados totales (Incluyendo duplicados): 27
Encontrada(s) 1 página(s)

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)

Microbiome Differential Abundance Methodologies to Detect Relevant Taxa Associated with Chemotherapy Toxicity Rate in Colorectal Cancer

  • Martin-De Arribas, E.
  • Conde-Pérez, Kelly
  • Aja-Macaya, Pablo
  • Vallejo, J. A.
  • Bou, Germán
  • López-Cheda, Ana
  • Jácome, M. A.
  • Ladra, Susana
  • Poza, Margarita
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG The data supporting this study are publicly available in the National Centre for Biotechnology Information (NCBI) Sequence Read Archive (SRA) database under the accession codes PRJNA911189 and PRJNA893853. The complete metadata set used in this paper is available in our GitHub repository (see Section 2.6). The external validation dataset used in this study will be available in the GitHub repository for this paper (see Section 2.6). At the same time, it can be downloaded directly from the original paper (Hillege et al. 2025). Supplementary material is available at Bioinformatics Advances online., [Abstract]: Motivation. The interplay between microbial communities and treatment outcomes represents a promising area in pharmacomicrobiomics. Identifying microbial biomarkers that differentiate toxicity levels could inform personalized cancer strategies. However, biomarker identification is strongly influenced by methodological choices in differential abundance analysis (DAA), and most studies focus on individual outcomes despite toxicity being inherently multifactorial. In this study, we defined a multi-dimensional toxicity variable integrating clinical symptoms and treatment modifications to stratify colorectal cancer patients. We then evaluated six widely used DAA methods (ALDEx2, ANCOM-BC, DESeq2, LEfSe, LinDA, and ZicoSeq) to assess how analytical variability affects the detection of microbiome signatures associated with chemotherapy-related toxicity. Analyses were performed under different preprocessing and multiple-testing correction strategies, and consistency was further examined using an independent validation dataset. Results. Substantial variability was observed across methods, with limited overlap in detected taxa but moderate concordance in effect-size rankings. ANCOM-BC showed the most consistent overall performance across analytical scenarios, although trade-offs remained between taxa detection, ranking, and direction of association. Despite this variability, a subset of taxa was consistently identified across methods, including Parvimonas, Eubacterium ventriosum group, and Ruminococcus in the low-toxicity group, and members of the Lachnospiraceae family, such as Fusicatenibacter, Lachnospira, and the Lachnospiraceae NK4A136 group, in the severe-toxicity group. Analyses in the external validation dataset supported the reproducibility of methodological patterns, despite differences in cohort composition and sequencing strategy. These findings highlight the methodological dependence of microbiome biomarker discovery and the potential of pre-treatment microbial signatures to stratify toxicity risk. View collectively, our results support a context-dependent approach to DAA method selection in clinical microbiome studies., This work was supported by the Spanish Ministry of Science and Innovation with funds from the European Union NextGenerationEU; GAIN and MRR funds, Recovery, Transformation and Resilience Plan (PRTRC17.I1) and from the Autonomous Community of Galicia within the framework of the Biotechnology Plan Applied to Health (DATAMICROCCR); GAIN/Xunta de Galicia (GRC: ED431C 2025/34, ED431C-2024/14); grant PID2023-147127OB-I00, funded by MCIN/ AEI/10.13039/501100011033/; and the Xunta de Galicia through the collaboration agreement between the Department of Culture, Education, Vocational Training and Universities and the Galician universities for the reinforcement of the research centres of the Galician University System, CIGUS (CITIC). This study also received funding from the Instituto de Salud Carlos III (ISCIII), Spain, through the projects PI20/00413 and PI23/00696, co-funded by the European Union [to M.P.]. Finally, project IN607A 2024/09, awarded to G.B., also funded this research. Funding for open access charge: Universidade da Coruña/CISUG., Xunta de Galicia; ED431C 2025/34, Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; IN607A 2024/09, https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/bioinformaticsadvances/6/1/10.1093_bioadv_vbag148/1/vbag148_supplementary_data.pdf?Expires=1786057871&Signature=3JyWGTBNJMk04Mo5YA478REuErHCckTX6BI6fB6LVUetTCLbgdQT2flCwWz3IJduY3I0wJGqrcGVq1LgHo5M3e8jAQggncSqnXBgVbCB~TCudFbcDFAKw3i8rF34P56E58L4krtm0F8cRDuW3WwGPZUvqxlETfuGcCUBU9XkLTLyOojjwVw78QXwxPN80aqHX9qNlKvwfcsVdRh52UpoK1rCRUHFQQrSGevPmMBL0M4tiCKg~CX7kUolD8o3BZvyKe4QzGXro7pEZYfRybU53ke-R3t0b05bo-oIK0xqkCwDqMhZZgriJYrleG1hpgW8EZ1Iyy3iwi7g4O86QmwrZQ__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA
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 Contribución de congreso (ConferenceOutput)

Presmoothed Latency Estimation of Cardiotoxicity in Breast Cancer Patients

  • Saavedra, S.
  • López-Cheda, Ana
  • Jácome, M. A.
Presentado en: VIII Congreso Xove TIC: impulsando el talento científico. Octubre, 2025, A Coruña., [Abstract] Breast cancer remains the leading cancer diagnosis among women. While treatment advances have significantly improved patient outcomes, some of the adverse effects associated with these therapies have been linked to cardiotoxicity, potentially compromising heart function. Understanding the time when such toxicity emerges is crucial to optimizing patient care and monitoring strategies. This study aims to estimate the time until cardiotoxicity appears, known as latency. The idea is to apply an innovative adaptation of traditional cure models considering a presmoothing method., This work, integrated into the framework of PERTE for Vanguard Health, has been co-financed by the Spanish Ministry of Science, Innovation and Universities with funds from the European Union Next Generation EU, 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. Besides, this work is part of the grant 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)., Xunta de Galicia; ED431G 2023/01, Xunta de Galicia; ED431C-2024/14
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)

A Bootstrap Goodness-of-fit Test for the Variogram

  • Fernández-Casal, Rubén
  • Castillo-Páez, Sergio
  • Francisco-Fernández, Mario
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG Three supplementary files accompany this article. The file SM_gof_svar.pdf contains the complete reproducible analysis of the real-data examples, including all R code, figures, and numerical results. The file precipgal.RData includes the precipitation data used in the Galicia example, and gof_svar.R provides the full R script for the real-data analysis. These materials allow the results to be fully reproduced., [Abstract]: This paper introduces a nonparametric bootstrap test for assessing the adequacy of parametric variogram models in geostatistical processes with a non-constant trend. Several goodness-of-fit procedures for variogram models have been developed under stationarity. In applied settings, however, the trend is often not constant and must be estimated prior to variogram analysis. A common practice is therefore to detrend the data and apply these procedures to the resulting residuals. This strategy induces systematic bias in the empirical variogram and compromises the validity of inference, particularly when the trend is estimated nonparametrically. To address this issue, we propose a bootstrap testing procedure that explicitly accounts for trend estimation by incorporating a bias-corrected variogram estimator obtained through an iterative local linear approach. The method accommodates both simple and composite null hypotheses for general parametric variogram models, with spatial independence arising as a special case. Unlike residual-based tests for independence or classical procedures for assessing specific parametric forms, the proposed approach corrects the bias induced by detrending and yields more reliable inference. Simulation studies show that the method achieves accurate calibration and substantially improved power across a wide range of dependence structures and trend configurations. Two real-data examples further illustrate its practical relevance when both the trend and the spatial dependence must be estimated from irregularly spaced observations. The methodology is implemented using readily available tools in the R environment, facilitating its use in applied spatial analysis., Funding for open access charge: Universidade da Coruña/CISUG. This work was supported by the Spanish Ministry of Science, Innovation and Universities under grant PID2023-147127OB-I00 (ERDF/EU, MCIN/AEI/10.13039/501100011033). Additional support was provided by the Xunta de Galicia through the project ED431C-2024/14 (Grupos de Referencia Competitiva), and by CITIC, a Research Center accredited for excellence within the Galician University System and member of the CIGUS Network, co-financed by the EU through the ERDF Galicia 2021–27 operational programme (Ref. ED431G-2023/01). The research of Sergio Castillo-Páez was also supported by the Department of Exact Sciences and the Doctoral Program in Administration of Universidad de las Fuerzas Armadas ESPE (Ecuador)., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G-2023/01
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 Tesis doctoral (DoctoralThesis)

Flexible Cure Models in Data Science to Predict Sustained Remission in Rheumatoid Arthritis

  • Monroy-Castillo, Blanca E.
[Abstract] Mixture cure models have become increasingly important in modern survival analysis, particularly in clinical contexts where a proportion of individuals may never experience the event of interest. These models allow researchers to distinguish between susceptible and long-term remission groups, offering deeper insight into longterm outcomes and covariate effects. Motivated by these advantages, this thesis contributes new methodological developments for assessing covariate significance in the cure proportion. A comparative study of two estimators of the distance correlation was conducted and an improved estimator that enhances stability and performance was introduced, especially in small or moderate samples. Building on this foundation, we propose four nonparametric hypothesis tests for evaluating covariate effects on the cure rate. Three of these tests rely on the martingale difference correlation, a conditional extension of the distance correlation that captures nonlinear and nonmonotonic relationships. Extensive simulation studies show that the new tests maintain the nominal significance level and achieve high power across a broad range of scenarios, outperforming existing nonparametric alternatives. The methodological contributions are implemented in a new R package, MDCcure, which provides functions for martingale difference correlation and all proposed tests. Finally, the methods are applied to a rheumatoid arthritis dataset to investigate factors associated with sustained remission under treatment optimization strategies., [Resumen] Los modelos de curación del tipo mixtura han adquirido una importancia creciente en el análisis de supervivencia moderno, especialmente en contextos clínicos donde una proporción de individuos puede no llegar nunca a experimentar el evento de interés. Estos modelos permiten diferenciar entre los individuos susceptibles y aquellos en remisión a largo plazo, proporcionando una comprensión más profunda de los resultados a largo plazo y de los efectos de las covariables. Motivada por estas ventajas, esta tesis aporta nuevos desarrollos metodológicos para evaluar la significancia de covariables en la proporción de curación. Se llevó a cabo un estudio comparativo de dos estimadores de la correlación de distancia y se introdujo un estimador mejorado que aumenta la estabilidad y el rendimiento, especialmente en muestras pequeñas o moderadas. Sobre esta base, proponemos cuatro pruebas de hipótesis no paramétricas para evaluar los efectos de covariables en la tasa de curación. Tres de estas pruebas se basan en la correlación de diferencia de martingalas (martingale difference correlation), una extensión condicional de la correlación de distancia (distance correlation) que captura relaciones no lineales y no monótonas. Amplios estudios de simulación muestran que las nuevas pruebas mantienen el nivel de significancia nominal y alcanzan una alta potencia en una amplia variedad de escenarios, superando a las alternativas no paramétricas existentes. Las contribuciones metodológicas se implementan en un nuevo paquete de R, MDCcure, que ofrece funciones para la correlación de diferencia de martingalas y todas las pruebas propuestas. Finalmente, los métodos se aplican a un conjunto de datos de artritis reumatoide para investigar los factores asociados con la remisión sostenida bajo estrategias de optimización del tratamiento., [Resumo] Os modelos de curación de tipo mestura converteronse nunha ferramenta cada vez máis importante na análise de supervivencia moderna, especialmente en contextos clínicos nos que unha proporción de individuos pode non experimentar nunca o acontecemento de interese. Estes modelos permítennos diferenciar entre individuos susceptibles e aqueles en remisión a longo prazo, proporcionando unha comprensión máis profunda dos resultados a longo prazo e dos efectos das covariables. Impulsada por estas vantaxes, esta tese contribúe con novos desenvolvementos metodolóxicos para avaliar a significación das covariables na proporción de curación. Realízase un estudo comparativo de dous estimadores de correlación de distancia e introdúcese un estimador mellorado que aumenta a estabilidade e o rendemento, especialmente en mostras pequenas ou moderadas. Sobre esta base, propoñemos catro probas de hipóteses non paramétricas para avaliar os efectos dos covariables na taxa de curación. Tres destas probas baséanse na correlación de diferenzas de martingala, unha extensión condicional da correlación de distancia que capta relacións non lineais e non monótonas. Amplos estudos de simulación amosan que as novas probas manteñen o nivel de significación nominal e acadan un poder alto nunha ampla variedade de escenarios, superando as alternativas non paramétricas existentes. As contribucións metodolóxicas impleméntanse nun novo paquete de R, MDCcure, que proporciona funcións para a correlación de diferenzas martingalas e todos os tests propostos. Finalmente, aplícanse os métodos a un conxunto de datos de artrite reumatoide para investigar factores asociados á remisión sostida baixo estratexias de optimización do tratamento., This research was supported by the International, Interdisciplinary and Intersectoral Information and Communications Technology PhD programme (3-i ICT) granted to CITIC and supported by the European Union through the Horizon 2020 research and innovation programme under a Marie Sklodowska-Curie agreement (H2020-MSCACOFUND). Besides, this work is part of the grant PID2023-147127OB-I00 ‘ERD-F/EU’, funded by MICIU/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)., Xunta de Galicia; ED431G 2023/01, Xunta de Galicia; ED431C-2024/14
RUC. Repositorio da Universidade da Coruña
Publicación Contribución de congreso (ConferenceOutput)

Presmoothed Latency Estimation of Cardiotoxicity in Breast Cancer Patients

  • Saavedra, S.
  • López-Cheda, Ana
  • Jácome, M. A.
Presentado en: VIII Congreso Xove TIC: impulsando el talento científico. Octubre, 2025, A Coruña., [Abstract] Breast cancer remains the leading cancer diagnosis among women. While treatment advances have significantly improved patient outcomes, some of the adverse effects associated with these therapies have been linked to cardiotoxicity, potentially compromising heart function. Understanding the time when such toxicity emerges is crucial to optimizing patient care and monitoring strategies. This study aims to estimate the time until cardiotoxicity appears, known as latency. The idea is to apply an innovative adaptation of traditional cure models considering a presmoothing method., This work, integrated into the framework of PERTE for Vanguard Health, has been co-financed by the Spanish Ministry of Science, Innovation and Universities with funds from the European Union Next Generation EU, 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. Besides, this work is part of the grant 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)., Xunta de Galicia; ED431G 2023/01, Xunta de Galicia; ED431C-2024/14
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

MDCcure: An R package for martingale difference correlation and hypothesis testing in mixture cure models

  • Monroy-Castillo, Blanca E.
  • Jácome, M. A.
  • Cao, Ricardo
[Abstract] Background: Understanding the relationship between covariates and clinical outcomes is a fundamental goal in biostatistics, particularly in the context of survival analysis and cure models. Traditional methods often lack the flexibility to assess complex dependencies or to evaluate covariate effects in a nonparametric manner. Moreover, there is a growing need for computationally efficient tools that integrate hypothesis testing and model diagnostics within the framework of long-term survival and cure probability estimation. Methods: We present a comprehensive R package that implements both novel and existing methods across three key areas: dependency analysis, nonparametric hypothesis testing in cure models, and goodness-of-fit testing for the cure rate in mixture cure models. For dependency analysis, we include functions based on martingale difference correlation and divergence to assess covariate effects on the conditional mean. To test the significance of covariates on cure probability, we propose a nonparametric framework comprising four approaches, three based on martingale difference correlation and one based on distance. An extended version of the test also allows for adjustment by a second covariate. For model checking, we introduce the goft() function for analytical goodness-of-fit testing and plotCure() for visual comparison of parametric and nonparametric cure probability estimates under logit, probit, and cloglog link functions. Results: The proposed methods showed a good performance in both simulated and real-world datasets. Martingale-based dependency measures effectively identified covariates influencing the conditional mean, while the nonparametric tests accurately detected covariate effects on cure probability. The extended test enhanced interpretability in multivariable settings. Goodness-of-fit analyses confirmed that the proposed tools correctly identified appropriate link functions, with visual outputs from plotCure() closely matching analytical results from goft(). Overall, the package enables accurate and efficient inference in the analysis of cure models., This research was supported by the International, Interdisciplinary and Intersectoral Information and Communications Technology PhD programme (3-i ICT) granted to CITIC and supported by the European Union through the Horizon 2020 research and innovation programme under a Marie Sklodowska-Curie agreement (H2020-MSCA-COFUND). Besides, this work is part of the grant PID2023-147127OB-I00 ‘ERD- F/EU’, funded by MICIU/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)., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Thermal and Mechanical Properties of Textiles as a Basis for Assessing Microfibre Release

  • Pereira, Mercedes
  • López-Beceiro, Jorge
  • Díaz-Díaz, Ana-María
  • Artiaga, Ramón
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, [Abstract] Although there are several studies regarding textile fibre release, there is little information about how thermal and mechanical properties of textiles can influence the process. In this study, TG, DSC and thermal and dynamic mechanical analysis are used to characterize the textiles. As the variables involved in textile fibre release are numerous and with different origins, statistical techniques of multivariate analysis are useful to reduce the dimensions of the problem, allowing the visualization of all variables together, considering their relationships. It also provides a statistical support to the study of fibre release, through which we found that the most strongly related parameters to fibre release are the nature of the fibres, the type of filament, continuous or staple, the elastic modulus and the crystallinity. Overall, this study provides a new understanding on how the thermal and mechanical properties of fabrics influence fibre shedding., The authors are thankful to Inditex and Technical Advice companies. Funding for open access charge: Universidade da Coruña/CISUG. This research has also been supported by the Ministerio de Ciencia e Innovación under grant PID2023-147127OB-I00., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This research was financially supported by Inditex.
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 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)

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 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)

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 Artículo científico (JournalArticle)

Goodness-of-Fit Tests for Parametric Circular Regression with Spatial Dependence: an Application to Wave Direction Modeling

  • Meilán-Vila, Andrea
  • Francisco-Fernández, Mario
  • Crujeiras-Casais, Rosa M.
[Abstract]: This paper introduces goodness-of-fit tests for parametric circular regression models with spatially correlated errors, motivated by the analysis of wave direction data in the Adriatic Sea. The proposed methodology confronts a parametric model with a local linear kernel estimator that accounts for both the circular nature of the response and the spatial dependence in the observed field. Two test statistics are introduced: one based on the raw discrepancy between the parametric and nonparametric fits, and another that incorporates a smoothed version of the parametric model to improve robustness. Under suitable regularity conditions, we derive the asymptotic distribution of the test statistic based on the smoothed parametric fit, providing theoretical support for its validity. Calibration under the null hypothesis is achieved via spatial bootstrap techniques specifically adapted to circular data. A comprehensive simulation study demonstrates that the proposed tests are well-calibrated across various dependence scenarios and exhibit strong power against a wide range of alternatives. Finally, the methodology is illustrated through a detailed application to Adriatic Sea wave direction data, where a physically motivated parametric model is evaluated using the proposed tests. The results reveal no evidence against the model, supporting its adequacy as a parsimonious representation of the underlying spatial structure., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. Funding for APC: Universidad Carlos III de Madrid (Agreement CRUE-Madroño 2026). Research by A. Meilán-Vila and M. Francisco-Fernández has been supported by the grant PID2023-147127OB-I00 (”ERDF/EU”), funded by MCIN/AEI/10.13039/501100011033. Additionally, research by M. Francisco-Fernández has also been supported by the Xunta de Galicia through the ”Grupos de Referencia Competitiva” program (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 support 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). Research by R. M. Crujeiras has been supported by MICINN (Grant PID2024-158017NB-I00 ) and by the Xunta de Galicia (Grupos de Referencia Competitiva, ED431C 2025/03), all of them through ERDF funds., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01, Xunta de Galicia; ED431C 2025/03, Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature
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)

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)

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)

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)

Improved nonparametric estimation of the cure rate in mixture cure models using presmoothing

  • Saavedra, S.
  • López-Cheda, Ana
  • Jácome, M. A.
This is an Accepted Manuscript version of the following article, accepted for publication in Journal of Nonparametric Statistics . It is deposited under the terms of the Creative Commons Attribution-NonCommercial License (http://creativecommons.org/licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited., [Abstract]: Survival analysis studies the time up to an event of interest assuming that all individuals will undergo the event. In some cases, the event may not happen for all subjects. This condition gives rise to cure models, where a subset of the individuals will experience the event and other part will never suffer it. Estimating the probability of experiencing the event is crucial in these situations. Presmoothing is a technique that improves the efficiency of nonparametric estimates in survival analysis and, therefore, in cure models. We propose to estimate the cure rate nonparametrically using presmoothing. Presmoothing depends on a smoothing parameter which must be chosen appropriately. To understand the effect of presmoothing on the estimation of the probability of cure, a simulation study was carried out. Finally, the methodology was applied to a dataset of breast cancer patients., This work, integrated into the framework of PERTE for Vanguard Health, has been co-financed by the Spanish Ministry of Science, Innovation and Universities with funds from the European Union Next Generation EU, 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. Besides, this work is part of the grant 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)., Xunta de Galicia; ED431C-2024/14, Xunta de Galicia; ED431G 2023/01
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)

AI Driven TTS Procedure to Predict Sailcloth Mechanical Behaviour for Wind Assisted Propulsion in Marine Navigation

  • Baamonde-González, J.
  • Álvarez García, Ana
  • Tarrío-Saavedra, Javier
  • González Taboada, María
  • López-Beceiro, Jorge
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, [Abstract]: The shipping industry is transitioning towards more sustainable propulsion strategies due to increasingly stringent greenhouse gas regulations, with wind-assisted propulsion emerging as a viable near-term solution to improve energy efficiency, reduce emissions and boost blue economy. However, the long-term mechanical performance of sail materials under combined environmental and mechanical degradation remains insufficiently understood. This study investigates the effects of marine exposure and mechanical abrasion on the dynamic and creep behaviour of Dacron sailcloth. Dynamic Mechanical Analysis (DMA) shows that marine ageing reduces the storage modulus by up to 17.7% (weft) and 11.2% (warp), while abrasion causes a reduction of 8.1%, with no significant changes in glass transition temperature. Short-term creep tests (30–100 °C) are extended using the Time–Temperature Superposition (TTS) principle to construct creep compliance master curves at 30 °C, revealing a clear increase in compliance for aged materials. A Generalised Additive Model (GAM) is applied to capture nonlinear behaviour and quantify the influence of predictors, showing excellent agreement (R² > 0.99 for individual curves; R² = 0.933 for the global model). The results indicate that creep compliance increases significantly with ageing, with marine exposure producing a vertical shift of the master curve and abrasion inducing a horizontal shift, suggesting increased chain mobility. Time (83.41%) and ageing level (10.63%) are identified as the dominant factors governing deformation, while fabric orientation confirms orthotropic behaviour, with lower compliance in the weft direction. The combination of TTS and GAM provides a robust framework for predicting long-term sailcloth performance within the validated time domain (∼10⁸ s), although extrapolation beyond this range should be treated with caution., Financing the Open Access Fee: Universidade da Coruña/CISUG. The following sail makers provided the Dacron sailcloth used in this paper free of charge: North Sails, Cuntis and Tartaruga Mar, Baiona (Spain). The research of Javier Tarrío-Saavedra has been supported by the Ministerio de Ciencia e Innovación grant 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 ED4[]C-2020-14 and ED431C 2024/014), and by the CITIC (Centro de Investigación en Tecnologías de la Información y Comunicación), 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; ED4[]C-2020-14, Xunta de Galicia; ED431C 2024/014, Xunta de Galicia; ED431G 2023/01
RUC. Repositorio da Universidade da Coruña
Publicación Artículo científico (JournalArticle)

Fatigue Behavior of Joule-cured and Oven-cured Carbon/Epoxy Laminates: a Generalized Additive Model Analysis

  • Vázquez, Laura
  • Naya, Salvador
  • Tarrío-Saavedra, Javier
  • Álvarez García, Ana
  • López-Beceiro, Jorge
  • Pereira, Mercedes
  • Francisco-Fernández, Mario
Financiado para publicación en acceso aberto: Universidade da Coruña/CISUG, [Abstract]: Joule (resistive) heating is a promising alternative to conventional oven curing for carbon/epoxy composites because it provides fiber interior heating and can reduce cycle time and energy consumption. From a fatigue perspective, however, the influence of the curing route on long-term mechanical performance is less understood. Here we compare the fatigue response (three-point bending load versus cycle count) of [90/0/90] laminates manufactured by Joule heating and by oven cure. Using a data-driven framework based on generalized additive models (GAMs), we analyze the effects of curing route, isothermal temperature and dwell time, ramp rate, displacement-controlled testing, and specimen orientation. Nonlinear degradation trends are captured by semiparametric GAMs fitted to 27 coupons, totaling 365,059 observations. The curing route emerges as the dominant driver of fatigue degradation: Oven-cured coupons start at higher load but lose capacity more rapidly, whereas Joule-cured coupons exhibit a smoother and more stable decline with cycling. Vertical orientation consistently yields higher loads, while within the explored Joule design space (110–130, 3–12 min) the effects of isothermal temperature, dwell time, and ramp rate are comparatively modest. These results support the viability of Joule heating as an alternative to conventional curing from the standpoint of fatigue behavior, in addition to its known gains in processing efficiency. The GAM approach identifies the key design parameters, quantifies their relative importance, and provides an interpretable tool for comparing curing programs and guiding the optimization of electrically cured composite structures., Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. Funding for open access charge: Universidade da Coruña / CISUG. This research was made possible by the support of the Campus Industrial de Ferrol (CIF) and the Centro de Investigación en Tecnoloxías Navais e Industriais (CITENI) of the University of A Coruña. The research of Laura S. Vázquez has been supported by the Axencia Galega de Innovación (Xunta de Galicia) through a Doutoramento Industrial grant (2022). This work has also been supported by grant PID2023-147127OB-I00 (“ERDF/EU”), funded by MCIN/AEI/10.13039/501100011033, by the Ministry for Digital Transformation and Civil Service under grant TSI-100925-2023-1, by the Xunta de Galicia through the “Grupos de Referencia Competitiva” program (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 support 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