In this paper, we propose a new method for nding similarity of efects based on quantile regression models. Clustering of effects curves (cec) techniques are applied to quantile regression coefficients, which are one-to- one functions of the order of the quantile. We adopt the quantile regression coefficients modeling (qrcm) framework to describe the functional form of the coefficient functions by means of parametric models. The proposed method can be utilized to cluster the effect of covariates with a univariate response variable, or to cluster a multivariate outcome. We report simulation results, comparing our approach with the existing techniques. The idea of combining cec with qrcm permits simplifying computation and interpretation of the results, and may improve the ability to identify clusters.We illustrate a variety of applications, highlighting the advantages and the usefulness of the described method.
Sottile G, A.G. (2018). Clusters of effects curves in quantile regression models. COMPUTATIONAL STATISTICS, 34(2), 551-569 [10.1007/s00180-018-0817-8].
Data di pubblicazione: | 2018 | |
Titolo: | Clusters of effects curves in quantile regression models | |
Autori: | ADELFIO, Giada (Corresponding) | |
Citazione: | Sottile G, A.G. (2018). Clusters of effects curves in quantile regression models. COMPUTATIONAL STATISTICS, 34(2), 551-569 [10.1007/s00180-018-0817-8]. | |
Rivista: | ||
Digital Object Identifier (DOI): | http://dx.doi.org/10.1007/s00180-018-0817-8 | |
Abstract: | In this paper, we propose a new method for nding similarity of efects based on quantile regression models. Clustering of effects curves (cec) techniques are applied to quantile regression coefficients, which are one-to- one functions of the order of the quantile. We adopt the quantile regression coefficients modeling (qrcm) framework to describe the functional form of the coefficient functions by means of parametric models. The proposed method can be utilized to cluster the effect of covariates with a univariate response variable, or to cluster a multivariate outcome. We report simulation results, comparing our approach with the existing techniques. The idea of combining cec with qrcm permits simplifying computation and interpretation of the results, and may improve the ability to identify clusters.We illustrate a variety of applications, highlighting the advantages and the usefulness of the described method. | |
Settore Scientifico Disciplinare: | Settore SECS-S/01 - Statistica | |
Appare nelle tipologie: | 1.01 Articolo in rivista |
File in questo prodotto:
File | Descrizione | Tipologia | Licenza | |
---|---|---|---|---|
Sottile-Adelfio2019_Article_ClustersOfEffectsCurvesInQuant.pdf | Versione Editoriale | Administrator Richiedi una copia |