Title: A method selection guidance in dr-package
Authors: Jae Keun Yoo - Ewha Womans University (Korea, South) [presenting]
Abstract: Sufficient dimension reduction (SDR) turns out to be a useful dimension reduction tool in high-dimensional regression analysis. The dr-package has been developed to implement the four most popular SDR methods. However, the package does not provide any clear guideline on which method should be used given a data. Since the four methods may provide dramatically different dimension reduction results, the selection in dr package should be problematic to statistical practitioners. A basis-adaptive selection algorithm is developed to relieve the issue. A basic idea is to select a SDR method to provide a highest correlation between the basis estimates obtained by the four classical SDR methods. Real data example and numerical studies confirm practical usefulness of the algorithm.