Research Abstracts Online
January 2010 - March 2011
University of Minnesota Twin Cities
College of Liberal Arts
PI: Xiaotong Shen
Group Pursuit Through a Regularization Solution Surface
Extracting grouping structure or identifying homogenous subgroups of predictors in regression is crucial for high-dimensional data analysis. When captured in a regression model, grouping enables researchers to enhance predictive performance and to facilitate a model’s interpretability. Grouping pursuit extracts homogenous subgroups of predictors most responsible for outcomes of a response.
This project explores gene network analysis, where grouping reveals gene functionalities with regard to progression of a disease. To address challenges in grouping pursuit, these researchers introduce a novel homotopy method for computing an entire solution surface through regularization involving a piecewise linear penalty. This non-convex and over-complete penalty permits adaptive grouping and nearly unbiased estimation, which is treated with a novel concept of grouped subdifferentials and difference convex programming for efficient computation.
Lingzhou Sue, Graduate Student
Meng-Hsuan Wu, Graduate Student
Changqing Ye, Graduate Student
Yiping Yuan, Graduate Student
Yunzhang Zhu, Graduate Student