Title: Supervised random projection T test for two sample test in high dimension
Authors: Radhendushka Srivastava - Indian Institute of Technology Bombay (India) [presenting]
Abstract: In high dimension, testing the equality of mean of two populations have been explored by several researchers. The RAndom Projection T Test (RAPTT) is an exact test for equality of means of two normal populations. The RAPTT does not require constraints on the dimension and the sample size. However, the random projection inserts additional randomness in the decision making. We use supervised random projection based on the data and propose Supervised RAndom Projection T test (SRAPTT). We illustrate the advantage of supervised through simulation and gene expression data.