Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
6868608 | Computational Statistics & Data Analysis | 2018 | 13 Pages |
Abstract
This article investigates the feature screening procedure for multivariate response varying coefficient linear models. A new conditional canonical correlation coefficient is proposed to characterize the correlation between each predictor and the multivariate response. It is shown that the proposed method is more powerful to distinguish the informative features from the noises than the existing competitors, especially for the case of high-dimensional response. The ranking consistency and the sure screening property are established for the new method. Meanwhile, an iterative version of the feature screening procedure is also introduced. Both the numerical simulations and real data analysis are conducted to illustrate the effectiveness of our method.
Related Topics
Physical Sciences and Engineering
Computer Science
Computational Theory and Mathematics
Authors
Jun Lu, Lu Lin,