A0214
Title: On high-dimensional data analysis
Authors: Wei-Cheng Hsiao - Soochow University (Taiwan) [presenting]
Ching-Kang Ing - National Tsing Hua University (Taiwan)
Abstract: Big data, ubiquitous in various fields of natural and social sciences, often contains a large number of variables and features. This motivates the study of the model (variable or feature) selection problems in high-dimensional sparse models. A novel high-dimensional model selection procedure is introduced and demonstrates how variable selection consistency in high-dimensional interaction models is achieved. In addition, motivated by wafer data, a new high-dimensional model identification method is proposed and its selection consistency is obtained in situations where the location and dispersion components of the model obey sparsity conditions. Simulations and real data analysis are given to illustrate the finite sample performance of the proposed methods.
