The Application of Targeted Variable Importance Measurement in Dimension Reduction in Gene Expression Data

• Activity Number: 124
• Type: Contributed
• Date/Time: Monday, August 1, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Hui Wang*+ and Mark van der Laan
• Companies: Stanford University and University of California at Berkeley
• Keywords: targeted maximum likelihood estimation ; semiparametric model ; variable importance measurement ; dimension reduction

Abstract

When a large number of candidate variables are present, a dimension reduction procedure is usually conducted to reduce the variable space before the subsequent analysis can be carried out. Inspired from the causal inference literature, we demonstrate that the variable importance measurement (VIM) based on targeted maximum likelihood estimation (TMLE) can be used for the purpose of dimension reduction. The TMLE-VIM is a two-stage procedure. The first stage resorts to a machine learning algorithm such as LARS and random forest. The second step improves the first stage estimation with respect to the variable of interest. Hence, TMLE-VIM enjoys the prediction power of machine learning algorithms, accounts for the correlation structures among variables, and at the same time produces more accurate variable rankings. When utilized in dimension reduction, TMLE-VIM can help to obtain the shortest possible list with the most truly associated variables.