Scholarlink

<p>We develop a novel feature selection method for linear regression with multivariate responses in ultrahigh-dimensional data analysis. This method is constructed under the framework of False Discovery Rate (FDR) control for multiple testing, and it employs a multiple data-splitting strategy. In each splitting, the data is divided into two disjoint parts. The first part is utilized for feature screening based on R-Vector (RV) correlation, and multiple testing is then conducted on the selected features for both parts. The z-values of the statistics are aggregated to control the FDR, and the set of important features is determined by rejecting the null hypotheses. The asymptotic theory of FDR control for this method is established under mild conditions. Additionally, we evaluate the finite sample performance of our feature selection procedure through Monte Carlo simulations. Finally, we apply this approach to detect important human genes associated with psychological well-being.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Feature screening via false discovery rate control for linear model with multivariate responses

  • 期刊论文

  • 发表日期：

    2025年1月23日

查看全文
  • Congran Yu,
  • Hengjian Cui

摘要

We develop a novel feature selection method for linear regression with multivariate responses in ultrahigh-dimensional data analysis. This method is constructed under the framework of False Discovery Rate (FDR) control for multiple testing, and it employs a multiple data-splitting strategy. In each splitting, the data is divided into two disjoint parts. The first part is utilized for feature screening based on R-Vector (RV) correlation, and multiple testing is then conducted on the selected features for both parts. The z-values of the statistics are aggregated to control the FDR, and the set of important features is determined by rejecting the null hypotheses. The asymptotic theory of FDR control for this method is established under mild conditions. Additionally, we evaluate the finite sample performance of our feature selection procedure through Monte Carlo simulations. Finally, we apply this approach to detect important human genes associated with psychological well-being.