A Theoretical Study of Multiple Testing Problem

• Main Authors: Shih-Yin Huang, 黃世吟
• Other Authors: Wen-Da Lo
• Format: Others
• Language: en_US
• Published: 2009
• Online Access: http://ndltd.ncl.edu.tw/handle/57067422581895266060

博士 === 國立中正大學 === 應用數學研究所 === 97 === Recently, there has been increased interest in the field of multiple testing due to a novel technology, such as gene expression arrays used to search for significant factors related to some kind of event such as disease. After, the multiple testing problem started to become practical and draw a lot of attention. For multiple testing, the major challenge is to design methods that while retaining reasonable power owing to various random situation. The aim of this paper is to construct optimal procedure of power for mutilple testing problem based on controlling TFDR. In particular, the TFDR procedure has an optimal property similar to Neyman - Pearson Theorem. The idea is an exchange rate of TPR/FPR similar to the power resulting from the unit error in the Neyman - Pearson Theorem. Besides, the comparisons with a similar result, ODP (Optimal Discovery Procedure) proposed by Storey, are also discussed in the article.