Skip to main content

NUEVOS AVANCES METODOLOGICOS Y COMPUTATIONALES EN ESTADISTICA NO PARAMETRICA Y SEMIPARAMETRICA

MTM2017-89422-P

•
Funding agency name Agencia Estatal de Investigación
Funding agency acronym AEI
Program Programa Estatal de Fomento de la Investigación Científica y Técnica de Excelencia
Subprogram Subprograma Estatal de Generación de Conocimiento
Call Proyectos de I+D (Excelencia / Generación de Conocimiento)
Call year 2017
State Plan 2017-2020 Plan Estatal de Investigación Científica y Técnica y de Innovación
Beneficiary centre UNIVERSIDAD DE VIGO

Resultados relacionados

Resultados totales (Incluyendo duplicados): 13
Encontrada(s) 1 página(s)

Minerva. Repositorio Institucional de la Universidad de Santiago de Compostela
Publication Artículo científico (JournalArticle)

Modeling conditional reference regions: application to glycemic markers

  • Lado Baleato, Óscar
  • Roca Pardiñas, Javier
  • Cadarso Suárez, Carmen María
  • Gude Sampedro, Francisco
Many clinical decisions are taken based on the results of continuous diagnostic tests. Usually, only the results of one single test is taken into consideration, the interpretation of which requires a reference range for the healthy population. However, the use of two different tests, can be necessary in the diagnosis of certain diseases. This obliges a bivariate reference region be available for their interpretation. It should also be remembered that reference regions may depend on patient variables (eg, age and sex) independent of the suspected disease. However, few proposals have been made regarding the statistical modeling of such reference regions, and those put forward have always assumed a Gaussian distribution, which can be rather restrictive. The present work describes a new statistical method that allows such reference regions to be estimated with no insistence on the results being normally distributed. The proposed method is based on a bivariate location-scale model that provides probabilistic regions covering a specific percentage of the bivariate data, dependent on certain covariates. The reference region is estimated nonparametrically and the nonlinear effects of continuous covariates via polynomial kernel smoothers in additive models. The bivariate model is estimated using a backfitting algorithm, and the optimal smoothing parameters of the kernel smoothers selected by cross-validation. The model performed satisfactorily in simulation studies under the assumption of non-Gaussian conditions. Finally, the proposed methodology was found to be useful in estimating a reference region for two continuous diagnostic tests for diabetes (fasting plasma glucose and glycated hemoglobin), taking into account the age of the patient
RODIN. Repositorio de Objetos de Docencia e Investigación de la Universidad de Cádiz
Publication Artículo científico (JournalArticle)

On sparse ensemble methods: An application to short-term predictions of the evolution of COVID-19

  • Benítez-Peña, Sandra
  • Carrizosa, Emilio
  • Guerrero, Vanesa
  • Jiménez-Gamero, M. Dolores
  • Martín-Barragán, Belén
  • Molero-Río, Cristina
  • Ramírez Cobo, Josefa
  • Romero Morales, Dolores
  • Sillero-Denamiel, M. Remedios
Since the seminal paper by Bates and Granger in 1969, a vast number of ensemble methods that combine different base regressors to generate a unique one have been proposed in the literature. The so-obtained regressor method may have better accuracy than its components, but at the same time it may overfit, it may be distorted by base regressors with low accuracy, and it may be too complex to understand and explain. This paper proposes and studies a novel Mathematical Optimization model to build a sparse ensemble, which trades off the accuracy of the ensemble and the number of base regressors used. The latter is controlled by means of a regularization term that penalizes regressors with a poor individual performance. Our approach is flexible to incorporate desirable properties one may have on the ensemble, such as controlling the performance of the ensemble in critical groups of records, or the costs associated with the base regressors involved in the ensemble. We illustrate our approach with real data sets arising in the COVID-19 context. (c) 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/ ), We thank the reviewers for their thorough comments and suggestions, which have been very valuable to strengthen the quality of the paper. This research has been financed in part by research projects EC H2020 MSCA RISE NeEDS (Grant agreement ID: 822214); FQM-329 and P18-FR-2369 (Junta de Andalucia, Spain); MTM2017-89422-P (Ministerio de Economia, Industria y Competitividad, Spain); PID2019-110886RB-I00 (Ministerio de Ciencia, Innovacin y Universidades, Spain); PR2019-029 (Universidad de Cadiz, Spain); PITUFLOW-CM-UC3M (Comunidad de Madrid and Universidad Carlos III de Madrid, Spain); and EP/R00370X/1 (EPSRC, United Kingdom). This support is gratefully acknowledged.
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Multiple comparison procedures for discrete uniform and homogeneous tests

  • Cousido Rocha, Marta
  • De Uña Alvarez, Jacobo
  • Döhler, Sebastian
Financiado para publicación en acceso aberto: Universidade de Vigo/CISUG, Discrete uniform and homogeneous p-values often arise in applications with multiple testing. For example, this occurs in genome wide association studies whenever a non-parametric one-sample (or two-sample) test is applied throughout the gene loci. In this paper, we consider multiple comparison procedures for such scenarios based on several existing estimators for the proportion of true null hypotheses, p0, which take the discreteness of the p-values into account. The theoretical guarantees of the several approaches with respect to the estimation of p0 and the false discovery rate control are reviewed. The performance of the discrete procedures is investigated through intensive Monte Carlo simulations considering both independent and dependent p-values. The methods are applied to three real data sets for illustration purposes too. Since the particular estimator of p0 used to compute the q-values may influence its performance, relative advantages and disadvantages of the reviewed procedures are discussed. Practical recommendations are given., Ministerio de Economía y Competitividad | Ref. BES-2015-074958, Ministerio de Economía y Competitividad | Ref. MTM2014-55966-P, Agencia Estatal de Investigación | Ref. MTM2017-89422-P, Xunta de Galicia | Ref. ED431C 2016/040
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Predicting pollution incidents through semiparametric quantile regression models

  • Roca Pardiñas, Javier
  • Ordóñez Galán, Celestino
In this paper we present a method to forecast pollution episodes using measurements of the pollutant concentration along time. Specifically, we use a backfitting algorithm with local polynomial kernel smoothers to estimate a semiparametric additive quantile regression model. We also propose a statistical hypothesis test to determine critical values, i.e., the values of the concentration that are significant to forecast the pollution episodes. This test is based on a wild bootstrap approach modified to suit asymmetric loss functions, as occurs in quantile regression. The validity of the method was checked with both simulated and real data, the latter from SO 2 emissions from a coal-fired power station located in Northern Spain., Ministerio de Economía y Competitividad | Ref. MTM2014-55966-P, Agencia Estatal de Investigación | Ref. MTM2017-89422-P
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Game-related statistics in the Spanish Water Polo League : differences between seasons, Estadísticas de partido en la Liga Española de Waterpolo : diferencias entre temporadas

  • Iglesias Pérez, Maria Carmen
  • García Ordoñez, Enrique
  • Touriño González, Carlos Francisco
The aim of this study was to identify the differences between seasons in meńs water polo regular competition by analysing the changes of game-related statistics. The sample comprised 88 games from the Spanish Professional Water Polo League (2011-2014). The game-related statistics were clustered into five groups: attacks in relation to the different playing situations, shots in relation to the different playing situations, attacks outcome, origin of shots and technical execution of shots. Univariate (ANOVA, Kruskal-Wallis and Generalized Linear Model tests) and multivariate (Discriminant) analyses were used to compare seasons, and statistical significance and effect sizes of the differences between seasons were calculated for the game-related statistics. The data were compositional data, therefore the variables were modified with additive log-ratio transformation. During the 2011-2012 season, the teams had significantly higher averages in even attacks (p<.01) and shots from zone 3 (p<.05). In the 2013-2014 season, teams had significantly higher averages in counterattack (p<.001), counterattack shots (p<.01), shots from zone 6 (p<.05) and 4 (p<.05). The variables that best distinguished between seasons were counterattack shots, counterattack, shots from zone 6 and 3, and even attacks. The group that best discriminated among seasons was «Origin of shots» (46.5% original sample and 42.3% cross-validation). The increase in counterattacks and in zone 6 shots seems to indicate a trend towards faster games and greater relevance of the centre forward. These findings can contribute to a better understanding of the evolution of water polo performance indicators, helping the coaches to prepare the players accordingly., El objetivo de este estudio fue identificar las diferencias entre temporadas para las estadísticas de juego en waterpolo masculino. La muestra consistió en 88 partidos de la Liga Española de Waterpolo (2011-2014). Las estadísticas de juego fueron agrupadas en cinco grupos: ataques según las diferentes situaciones de juego, lanzamientos según las situaciones de juego, resultado del ataque, origen de lanzamiento y ejecución técnica de lanzamiento. Se aplicaron varios métodos estadísticos, univariantes (ANOVA, Kruskal-Wallis y Modelo Lineal Generalizado) y multivariantes (análisis discriminante) para comparar las estadísticas entre las diferentes temporadas, calculándose diferencias significativas y tamaños del efecto. Debido a la existencia de datos composicionales, las variables fueron transformadas utilizando la función log-cociente aditiva. Se observó que durante la temporada 2011-2012 los equipos realizaron más ataques en igualdad (p<.01) y lanzamientos desde zona 3 (p<.05). En la temporada 2013-2014, realizaron más contraataques (p<.001), lanzamientos en contraataque (p<.01) y lanzamientos desde zona 6 (p<.05) y 4 (p<.05). Las variables que mejor discriminaron entre temporadas fueron los contrataques, lanzamientos en contrataque, lanzamientos de zona 6 y 3, y ataques en igualdad. El grupo que mejor discriminó entre temporadas fue «Origen de lanzamiento» (46,5% muestra original y 42,3% validación cruzada). El aumento de contraataques y lanzamientos de zona 6 parecen indicar una tendencia hacia juegos más rápidos y mayor relevancia del boya. Estos hallazgos pueden contribuir a mejorar el conocimiento de la evolución de los indicadores de rendimiento en waterpolo, ayudando a los entrenadores para preparar a sus jugadores de acuerdo a estos cambios., Ministerio de Economía y Competitividad | Ref. MTM2014-55966-P, Xunta de Galicia | Ref. ED431C 2016/040, Agencia Estatal de Investigación | Ref. MTM2017-89422-P, Xunta de Galicia | Ref. ED431G/02
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

clustcurv: an R package for determining groups in multiple curves

  • Martinez Villanueva, Nora
  • Sestelo Perez, Marta
  • Meira Machado, Luis
  • Roca Pardiñas, Javier
In many situations, it could be interesting to ascertain whether groups of curves can be performed, especially when confronted with a considerable number of curves. This paper introduces an R package, known as clustcurv, for determining clusters of curves with an automatic selection of their number. The package can be used for determining groups in multiple survival curves as well as for multiple regression curves. Moreover, it can be used with large numbers of curves. An illustration of the use of clustcurv is provided, using both real data examples and artificial data., Agencia Estatal de Investigación | Ref. MTM2017-89422-P, Agencia Estatal de Investigación | Ref. MTM2017-82379-R
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Selecting the number of categories of the lymph node ratio in cancer research: A bootstrap-based hypothesis test

  • Barrio Beraza, Irantzu
  • Roca Pardiñas, Javier
  • Arostegui Madariaga, Inmaculada
The high impact of the lymph node ratio as a prognostic factor is widely established in colorectal cancer, and is being used as a categorized predictor variable in several studies. However, the cut-off points as well as the number of categories considered differ considerably in the literature. Motivated by the need to obtain the best categorization of the lymph node ratio as a predictor of mortality in colorectal cancer patients, we propose a method to select the best number of categories for a continuous variable in a logistic regression framework. Thus, to this end, we propose a bootstrap-based hypothesis test, together with a new estimation algorithm for the optimal location of the cut-off points called BackAddFor, which is an updated version of the previously proposed AddFor algorithm. The performance of the hypothesis test was evaluated by means of a simulation study, under different scenarios, yielding type I errors close to the nominal errors and good power values whenever a meaningful difference in terms of prediction ability existed. Finally, the methodology proposed was applied to the CCR-CARESS study where the lymph node ratio was included as a predictor of five-year mortality, resulting in the selection of three categories., Gobierno Vasco | Ref. IT1294-19, Gobierno Vasco | Ref. KK-2020/00049, Agencia Estatal de Investigación | Ref. SEV-2017-0718, Ministerio de Economía y Competitividad | Ref. MTM2014-55966-P, Ministerio de Economía y Competitividad | Ref. MTM2016-74931-P, Agencia Estatal de Investigación | Ref. MTM2017-89422-P, Xunta de Galicia | Ref. ED431G2019/06, Instituto de Salud Carlos III
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Network intrusion detection system for DDoS attacks in ICS using deep autoencoders

  • Ortega Fernández, Inés
  • Sestelo Perez, Marta
  • Burguillo Rial, Juan Carlos
  • Piñón Blanco, Camilo
Anomaly detection in industrial control and cyber-physical systems has gained much attention over the past years due to the increasing modernisation and exposure of industrial environments. Current dangers to the connected industry include the theft of industrial intellectual property, denial of service, or the compromise of cloud components; all of which might result in a cyber-attack across the operational network. However, most scientific work employs device logs, which necessitate substantial understanding and preprocessing before they can be used in anomaly detection. In this paper, we propose a network intrusion detection system (NIDS) architecture based on a deep autoencoder trained on network flow data, which has the advantage of not requiring prior knowledge of the network topology or its underlying architecture. Experimental results show that the proposed model can detect anomalies, caused by distributed denial of service attacks, providing a high detection rate and low false alarms, outperforming the state-of-the-art and a baseline model in an unsupervised learning environment. Furthermore, the deep autoencoder model can detect abnormal behaviour in legitimate devices after an attack. We also demonstrate the suitability of the proposed NIDS in a real industrial plant from the alimentary sector, analysing the false positive rate and the viability of the data generation, filtering and preprocessing procedure for a near real time scenario. The suggested NIDS architecture is a low-cost solution that uses only fifteen network-based features, requires minimal processing, operates in unsupervised mode, and is straightforward to deploy in real-world scenarios., Axencia Galega de Innovación | Ref. IN854A 2019/15, Centro para el Desarrollo Tecnológico Industrial | Ref. CER-20191012, Agencia Estatal de Investigación | Ref. MTM2017-89422-P, Financiado para publicación en acceso aberto: Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

An improved statistical approach for conformational analysis from 13C NMR chemical shifts

  • Cotos Yáñez, Tomas Raimundo
  • Pérez González, Ana
  • Iglesias Antelo, Maria Beatriz
  • Muñoz López, Luis
Financiado para publicación en acceso aberto: Universidade de Vigo/CISUG, In this article we propose an improved statistical approach for the estimation of the mole fraction of each conformer in a given conformational equilibrium from the measured averaged 13C NMR chemical shifts. As each compound has a different number of conformers, an iterative process of fitting (backfitting type) is necessary. In each iteration, two estimates are made, i) fitting to a ridge regression model with equality and inequality constraints and selection of variables for each compound and ii) re-fitting for all compounds. This new procedure leads to consistent results, with low standard error and good p-values, allowing the estimation of the conformational composition for small molecules with a reduced number of conformers., Ministerio de Economía y Competitividad | Ref. SAF2013-48397-R, Ministerio de Economía y Competitividad | Ref. SAF2016-77620-R, Ministerio de Economía y Competividad | Ref. MTM2016-76969-P, Agencia Estatal de Investigación | Ref. MTM2017-89422-P, Agencia Estatal de Investigación | Ref. CTQ2017-85919-R
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Tesis doctoral (DoctoralThesis)

New contributions to the identification of groups in non parametric curves, Novas contribucións á identificación de grupos en curvas non paramétricas, Nuevas contribuciones a la identificación de grupos en curvas no paramétricas

  • Martinez Villanueva, Nora
Several problems in statistical inference can be considered as part of statistical modelling. The way to aboard inference problems leads to two different focus: parametric and nonparametric ones. Parametric models require assumptions about structure and distributional form, unlike nonparametric models in which the model structure is not specified in advance but is instead determined from data. As nonparametric models are flexible, their applicability has gained increasing importance in widely diverse fields, such as medicine, biology end engineering and their development continue at an active pace today. One of the main goals in nonparametric statistical modelling is to understand the dependence between a response variable and an explanatory variable through regression models. The comparison of these regression models has been widely investigated in the literature. This problem has also been abroad in many longitudinal medical studies in which is interesting to compare survival between patients. However, when the null hypothesis of equality of curves is rejected, ascertaining whether groups of curves can exist is still a challenge. Thus, the main objective of this dissertation is to determine groups of multiple survival curves and regression curves. Regarding this, a new procedure is presented. Its performance is analysed through simulation studies and it is proved that is able to assign curves to their corresponding groups and to select the optimal number of them. Additionally, the method is illustrated using different real data applications. Finally, with the aim of making these methods available for their usage in real life problems by researchers, developers or stakeholders, an R package (clustcurv v2.0.1) has been developed which is freely available from the CRAN (Comprehensive R Archive Network)., Varios problemas de inferencia estadística pueden considerarse como parte del modelado estadístico. La forma de abordar los problemas de inferencia lleva a dos enfoques diferentes: paramétricos y no paramétricos. Los modelos paramétricos requieren suposiciones sobre la estructura y la forma de distribución, a diferencia de los modelos no paramétricos en los que la estructura del modelo no se especifica de antemano, sino que se determina a partir de los datos. Como los modelos no paramétricos son flexibles, su aplicabilidad ha ganado una importancia cada vez mayor en campos muy diversos, como la medicina, la biología y la ingeniería y su desarrollo continúan a un ritmo activo en la actualidad. Uno de los principales objetivos del modelado estadístico no paramétrico es comprender la dependencia entre una variable de respuesta y una variable explicativa a través de modelos de regresión. La comparación de estos modelos de regresión ha sido ampliamente investigada en la literatura. Este problema también ha estado en el exterior en muchos estudios médicos longitudinales en los que es interesante comparar la supervivencia entre pacientes. Sin embargo, cuando se rechaza la hipótesis nula de igualdad de curvas, determinar si pueden existir grupos de curvas sigue siendo un desafío. Así, el principal objetivo de esta tesis es determinar grupos de múltiples curvas de supervivencia y curvas de regresión. Al respecto, se presenta un nuevo procedimiento. Se analiza su desempeño mediante estudios de simulación y se comprueba que es capaz de asignar curvas a sus grupos correspondientes y seleccionar el número óptimo de ellas. Además, el método se ilustra utilizando diferentes aplicaciones de datos reales. Finalmente, con el objetivo de hacer que estos métodos estén disponibles para su uso en problemas de la vida real por parte de investigadores, desarrolladores o partes interesadas, se ha desarrollado un paquete R (clustcurv v2.0.1) que está disponible gratuitamente en CRAN (Comprehensive R Archive Network), Varios problemas de inferencia estatística poden considerarse como parte do modelado estatístico. O xeito de abordar os problemas de inferencia leva a dous enfoques diferentes: paramétricos e non paramétricos. Os modelos paramétricos requiren suposicións sobre a estrutura e a forma da distribución, a diferenza dos modelos non paramétricos nos que a estrutura do modelo non se especifica previamente, senón que se determina a partir dos datos. Como os modelos non paramétricos son flexibles, a súa aplicabilidade cobrou cada vez máis importancia en campos moi diversos como a medicina, a bioloxía e a enxeñaría, e o seu desenvolvemento continúa a un ritmo activo na actualidade. Un dos obxectivos principais do modelado estatístico non paramétrico é comprender a dependencia entre unha variable de resposta e unha variable explicativa a través de modelos de regresión. A comparación destes modelos de regresión foi amplamente investigada na literatura. Este problema tamén estivo no exterior en moitos estudos médicos lonxitudinais nos que é interesante comparar a supervivencia entre pacientes. Non obstante, cando se rexeita a hipótese nula de igualdade de curvas, determinar se poden existir grupos de curvas segue sendo un desafío. Así, o obxectivo principal desta tese é determinar grupos de curvas de supervivencia múltiples e curvas de regresión. A este respecto, preséntase un novo procedemento. O seu rendemento analízase mediante estudos de simulación e compróbase que é capaz de asignar curvas aos seus grupos correspondentes e seleccionar o número óptimo deles. Ademais, o método está ilustrado usando diferentes aplicacións de datos reais. Finalmente, co obxectivo de facer estes métodos dispoñibles para o seu uso en problemas da vida real por investigadores, desenvolvedores ou partes interesadas, desenvolveuse un paquete R (clustcurv v2.0.1) que está dispoñible gratuitamente en CRAN (Comprehensive R Archive Network), Agencia Estatal de Investigación | Ref. MTM2017-89422-P
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Spatial heterogeneity in Spain for senior travel behavior

  • Losada Sánchez, María Nieves
  • Alén González, María Elisa
  • Cotos Yáñez, Tomas Raimundo
  • Domínguez Vila, Trinidad
Travel patterns are becoming more differentiated, influenced by new variables resulting from changes in lifestyle. The relevance of the senior segment for this industry, with the continuous aging population and their economic status, made this segment a very attractive group for the sector, and even more so in a country that is characterized by its aging, such as Spain. Spatial effects are considered a key element to understand this process, but there are only a few studies focusing on cross-cultural influences and the neighborhood context. For this purpose, the technique of geographically weighted principal component analysis (GWPCA) is applied in a novel way for the sector, showing different behavior patterns according to area of origin. The GWPCA is a localized version of the principal component analysis (PCA) used when there is a certain spatial heterogeneity in the structure of a multivariate data set. The results confirmed that GWPCA is an effective statistical methodology to research spatial heterogeneity for travel behavior, with clearly differentiated scenarios for the north, center and south of Spain, where the most determining factors in each case were hygiene and cleanliness, medical coverage and transport facilities., Agencia Estatal de Investigación | Ref. MTM2017-89422-P, Xunta de Galicia | Ref. ED431C 2016/040
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

A new semantic-based feature selection method for spam filtering

  • Méndez Reboredo, José Ramón
  • Cotos Yáñez, Tomas Raimundo
  • Ruano Ordás, David Alfonso
The Internet emerged as a powerful infrastructure for the worldwide communication and interaction of people. Some unethical uses of this technology (for instance spam or viruses) generated challenges in the development of mechanisms to guarantee an affordable and secure experience concerning its usage. This study deals with the massive delivery of unwanted content or advertising campaigns without the accordance of target users (also known as spam). Currently, words (tokens) are selected by using feature selection schemes; they are then used to create feature vectors for training different Machine Learning (ML) approaches. This study introduces a new feature selection method able to take advantage of a semantic ontology to group words into topics and use them to build feature vectors. To this end, we have compared the performance of nine well-known ML approaches in conjunction with (i) Information Gain, the most popular feature selection method in the spam-filtering domain and (ii) Latent Dirichlet Allocation, a generative statistical model that allows sets of observations to be explained by unobserved groups that describe why some parts of the data are similar, and (iii) our semantic-based feature selection proposal. Results have shown the suitability and additional benefits of topic-driven methods to develop and deploy high-performance spam filters., Xunta de Galicia | Ref. ED481B 2017/018, Xunta de Galicia | Ref. ED431C 2016-040, Agencia Estatal de Investigación | Ref. MTM2017-89422-P, Agencia Estatal de Investigación | Ref. TIN2017-84658-C2-1-R
RUA. Repositorio Institucional de la Universidad de Alicante
Publication Otros (Other)

New stochastic comparisons based on tail value at risk measures

  • Belzunce, Félix
  • Franco-Pereira, Alba M.
  • Mulero, Julio
  • Mulero, Julio|||/items/fd8c6a50-6e7e-45e6-8acf-e8b35316c5be
In this article we provide a new criterion for the comparison of claims, when we have conditional claims arising in stop loss contracts or contracts with franchise deductible. These stochastic comparisons are made on the basis of the Tail Value at Risk (also known as conditional tail expectation), just for a fixed level and beyond. In particular, we explain the interest of comparing these quantities, study some preservation properties and, in addition, we provide sufficient conditions for its study. Finally we illustrate its usefulness with some examples.