Graphical models via joint quantile regression with component selection

JOURNAL OF MULTIVARIATE ANALYSIS

卷:152

Article

Chun, Hyonho¹  Lee, Myung Hee²  Fleet, James C.³  Oh, Ji Hwan¹ 

[1] Purdue Univ, Dept Stat, W Lafayette, IN 47906 USA

[2] Weill Cornell Med, Dept Med, Ctr Global Hlth, New York, NY 10065 USA

[3] Purdue Univ, Dept Nutr Sci, W Lafayette, IN 47906 USA

关键词: Gaussian graphical models;    Conditional quantile;    Component selection;    Generalized additive models;   

DOI  :  10.1016/j.jmva.2016.07.012

来源: Elsevier

【 摘 要 】

A graphical model is used for describing interrelationships among multiple variables. In many cases, the multivariate Gaussian assumption is made partly for its simplicity but the assumption is hardly met in actual applications. In order to avoid dependence on a rather strong assumption, we propose to infer the graphical model via joint quantile regression with component selection, since the components of quantile regression carry information to infer the conditional independence. We demonstrate the advantages of our approach using simulation studies and apply our method to an interesting real biological dataset, where the dependence structure is highly complex. (C) 2016 Elsevier Inc. All rights reserved.

【 授权许可】

Free