Incorporation of sparsity information in large-scale multiple two-sample t-tests
报告人：刘卫东，上海交通大学教授，基金委优秀青年基金获得者，教育部新世纪人才
报告时间：10:30-11::30am, 12月19日，2014年
报告地点：数学楼二楼报告厅
报告摘要：Large-scale multiple two-sample Student's t testing problems often arise from the statistical analysis of scientic data. To detect components with di
erent values between two mean vectors, a well-known procedure is to apply the Benjamini and Hochberg (B-H) method and two-sample Student's t statistics to control the false discovery rate (FDR). In many applications, mean vectors are expected to be sparse or asymptotically sparse. When dealing with such type of data, can we gain
erent values between two mean vectors, a well-known procedure is to apply the Benjamini and Hochberg (B-H) method and two-sample Student's t statistics to control the false discovery rate (FDR). In many applications, mean vectors are expected to be sparse or asymptotically sparse. When dealing with such type of data, can we gain
more power than the standard procedure such as the B-H method with Student's t
statistics while keeping the FDR under control? The answer is positive. By exploiting the possible sparsity information in mean vectors, we present an uncorrelated
screening-based (US) FDR control procedure, which is shown to be more powerful
than the B-H method. The US testing procedure depends on a novel construction
of screening statistics, which are asymptotically uncorrelated with two-sample Student's t statistics. The US testing procedure is di
erent from some existing testing following screening methods (Reiner, et al., 2007; Yekutieli, 2008) in which independence between screening and testing is crucial to control the FDR, while the independence often requires additional data or splitting of samples. An inadequate splitting of samples may result in a loss rather than an improvement of statistical power. Instead, the uncorrelated screening US is based on the original data and does not need to split the samples. Theoretical results show that the US testing procedure controls the desired FDR asymptotically. Numerical studies are conducted and indicate that the proposed procedure works quite well.
erent from some existing testing following screening methods (Reiner, et al., 2007; Yekutieli, 2008) in which independence between screening and testing is crucial to control the FDR, while the independence often requires additional data or splitting of samples. An inadequate splitting of samples may result in a loss rather than an improvement of statistical power. Instead, the uncorrelated screening US is based on the original data and does not need to split the samples. Theoretical results show that the US testing procedure controls the desired FDR asymptotically. Numerical studies are conducted and indicate that the proposed procedure works quite well.