English

Sufficient dimension reduction with additional information

Methodology 2014-10-15 v1

Abstract

Sufficient dimension reduction is widely applied to help model building between the response YY and covariate XX. While the target of interest is the relationship between (Y,X)(Y,X), in some applications we also collect additional variable WW that is strongly correlated with YY. From a statistical point of view, making inference about (Y,X)(Y,X) without using WW will lose efficiency. However, it is not trivial to incorporate the information of WW to infer (Y,X)(Y,X). In this article, we propose a two-stage dimension reduction method for (Y,X)(Y,X), that is able to utilize the additional information from WW. The main idea is to confine the searching space, by constructing an envelope subspace for the target of interest. In the analysis of breast cancer data, the risk score constructed from the two-stage method can well separate patients with different survival experiences. In the Pima data, the two-stage method requires fewer components to infer the diabetes status, while achieving higher classification accuracy than conventional method.

Keywords

Cite

@article{arxiv.1410.3561,
  title  = {Sufficient dimension reduction with additional information},
  author = {Hung Hung and Chih-Yen Liu and Henry Horng-Shing Lu},
  journal= {arXiv preprint arXiv:1410.3561},
  year   = {2014}
}

Comments

26 pages, 4 figures, 1 table

R2 v1 2026-06-22T06:22:24.548Z