Robust Inference with Nonparametric Bayesian Models

• Activity Number: 61
• Type: Topic Contributed
• Date/Time: Sunday, August 4, 2013 : 4:00 PM to 5:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Steven MacEachern*+
• Companies: Ohio State University
• Keywords: Dirichlet process ; Huber loss ; mixture model

Abstract

Nonparametric Bayesian models, such as those based on the Dirichlet process or its many variants, provide a flexible class of models that allow us to fit widely varying patterns in data. Typical uses of the models include relatively low-dimensional driving terms to capture global features of the data along with a nonparametric structure to capture local features. The models are particularly good at handling outliers, a common form of local behavior, and examination of the posterior often shows that a portion of the model is chasing the outliers. The impact of the outliers carries over to the predictive distribution. This suggests the need for robust inference to discount the impact of the outliers on the overall analysis. This talk will illustrate strategies for making robust inference with these models.