Nonparametric Multivariate Mixture Model with Conditional Independence Assumption

• Activity Number: 251
• Type: Contributed
• Date/Time: Monday, August 4, 2014 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistical Learning and Data Mining
• Author(s): Xiaotian Zhu*+
• Companies: Penn State
• Keywords: mixture model ; nonparametric estimation ; independent component analysis ; Kullback-Leibler ; R package

Abstract

For the estimation of nonparametric multivariate mixture with conditional independence assumption, a new formulation of the objective function in terms of penalized smoothed Kullback-Leibler distance is proposed. The nonlinear majorization-minimization smoothing algorithm (NMMS) is derived from this perspective. Improved monotonicity property of the algorithm is dicovered and the existence of a solution to the main optimization problem is proved. Extension of the model and algorithm to incorporate independent component analysis (ICA) is presented. An R package that impelments the algorithms for real applications is also developed.