學術報告

報告人：林巍 副教授 （俄亥俄大學）

Title: On Selective Combination for Dimension Reduction

Abstract:

  Dimension reduction fo1r regression analysis has been one of the most popular topics in the past two decades.Thestudy in this area sees much progress with the introduction of the inverse regression method pioneered by Li (1991). Many of these methods are centered around a matrix, called the central matrix, which is then used to estimate the so-called central subspace. Although there are numerous proposals for the central matrices, none of them stands out in all cases. Thus, for a given data set, it remains unclear which of the existing central matrices should one use. Lots of efforts were  made to combine the benets of different central matrices, but they are either difficult to implement, or not completely data-driven, or are inconsistent in performance. In this work, we propose a simple procedure to selectively combine existing central matrices based an identification of the situation as either SIR-friendly or potentially SIR-unfriendly. We also introduce a new BIC (Bayesian information criterion) criterion as well as a selective algorithm to estimate the structural dimension of the central subspace. An extensive simulation study shows that our proposal works very favorably against other popular competitors.

報告時間：2016年4月28日（星期四）下午15:30

報告地點：科技樓南602