A Note on Sufficient Dimension Reduction
In this paper, we presented a theoretical result and then discussed possible applications of our result to SDR problems. In addition to providing insights into existing SDR methods when Y is univariate; our theorem also applies to multivariate responses, especially when the response takes the form of (Y,W), where Y is a continuous variable and W is categorical.
X. M. Wen, "A Note on Sufficient Dimension Reduction," Statistics & Probability Letters, Elsevier, Jan 2007.
The definitive version is available at http://dx.doi.org/10.1016/j.spl.2006.12.016
Mathematics and Statistics
Keywords and Phrases
censoring regression; multivariate dimension reduction; slicing; sufficient dimension reduction
Article - Journal
© 2007 Elsevier, All rights reserved.