|
Activity Number:
|
318
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistical Learning and Data Mining
|
| Abstract - #304362 |
|
Title:
|
Statistical Inference for High-Dimensional Data
|
|
Author(s):
|
Yingli Qin*+
|
|
Companies:
|
Iowa State University
|
|
Address:
|
428 S Walnut Ave, Ames, IA, 50010,
|
|
Keywords:
|
High dimensional ; large p small n ; Asymptotic Normality ; Gene set
|
|
Abstract:
|
We proposed a two sample test for means of high dimensional data when the data dimension is much larger than the sample size. The proposed test does not require any condition on the relationship between the data dimension and sample size explicitly. This offers much flexibility in analyzing high dimensional data. Researchers are also interested in the differences between the multivariate distributions across different treatments. By following the same path, I proposed a two sample test for distribution functions of high dimensional data. Asymptotic normality has been established and simulation results confirmed good performance of the proposed test statistic. An important application of the proposed test is in testing distribution functions for sets of genes in genetic studies. We demonstrate how to apply the test to gene sets for a Leukemia data set.
|
