Regularized Gaussian Mixture Modeling with Adaptive Covariance Shrinkage

• Activity Number: 284
• Type: Topic Contributed
• Date/Time: Tuesday, August 3, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Biometrics Section
• Author(s): Hyang Min Lee*+ and Jia Li
• Companies: Penn State and Penn State
• Keywords: Gaussian mixture models ; Covariance shrinkage ; EM algorithm ; , BIC-type penalized log-likelihood ; mixture model based clustering ; Model regularization

Abstract

We introduce a covariance shrinkage method for Gaussian mixture models that allows different components to have different levels of complexity. A complexity parameter is assigned to each component to determine the extent of shrinkage towards a diagonal or common covariance matrix. A BIC-type penalized log-likelihood is proposed to estimate the model parameters and the complexity parameters. A generalized EM algorithm is developed for model estimation. Based on both simulated and real data sets, we will compare the proposed covariance shrinkage method in terms of likelihood and parameter accuracy with covariance shrinkage using a single complexity parameter and estimation without shrinkage. We will also investigate the impact on clustering based on this new mixture modeling technique.