报告时间：2016年7月6日（周三）下午2:30

报告地点：京师学堂第七会议室

报 告 人：李根,哥伦比亚大学公共卫生学院生物统计系

报告题目：Supervised Integrated Factor Analysis for Multi-View Data

报告摘要：High dimensional data pose a great challenge to statistics, commonly known as the curse of dimensionality. Dimension reduction is a popular exploratory tool for processing high dimensional data to facilitate further statistical analyses. However, standard dimension reduction methods such as principal component analysis and factor analysis only apply to a single data set, and do not easily accommodate auxiliary supervision or any special features of data, such as functionality, sparsity, and multi-modality. In this talk, I will introduce a new supervised dimension reduction framework that is adapted to deal with modern data types. In particular, it can flexibly account for any auxiliary information in the reduction of primary data, incorporate variable selection and smoothing, and jointly decompose multiple related data sets when applicable. The proposed methods have broad applications in practice. I will demonstrate the advantage over existing methods using examples from bioinformatics and business analytics.