【主题】Simultaneous variable and factor selection via sparse group lasso in factor analysis

【报告人】王晴, 助教授

Wellesley College in Massachusetts, USA

【时间】2019年7月15日（星期一）15:30-16:30

【地点】上海财经大学统计与管理学院大楼1208会议室

【摘要】This talk considers variable and factor selection in factor analysis. We treat the factor loadings for each observable variable as a group, and introduce a weighted sparse group lasso penalty to the complete log-likelihood. The proposal simultaneously selects observable variables and latent factors of a factor analysis model in a data-driven fashion; it produces a more flexible and sparse factor loading structure than existing methods. For parameter estimation, we derive an expectation-maximization algorithm that optimizes the penalized log-likelihood. The tuning parameters of the procedure are selected by a likelihood cross-validation criterion that yieldssatisfactory results in various simulation settings. Simulation results reveal that the proposed method can better identify the possibly sparse structure of the true factor loading matrix with higher estimation accuracy than existing methods. A real data example is also presented to demonstrate its performance in practice.

【嘉宾简介】Qing (Wendy) Wang is an Assistant Professor of Statistics at Wellesley College in Massachusetts, USA. She received her Ph.D. in statistics, under the supervision of Dr. Bruce Lindsay, from the Pennsylvania State University in 2012. Prior to joining Wellesley, she was an Assistant Professor of Statistics at Williams College from 2012 to 2015 and at Bentley University in 2015-2016. Her research examines topics in U-statistics, nonparametric kernel density estimation, risk estimation, variance estimation, and cross-validation. In addition, she is also interested in penalization methods, model assessment tools, and multi-label classification. 

【主持人】王绍立