Bayesian Model Assessment in Factor Analysis with Incomplete Data

• Activity Number: 433
• Type: Contributed
• Date/Time: Tuesday, August 6, 2013 : 2:00 PM to 3:50 PM
• Sponsor: Section on Bayesian Statistical Science
• Author(s): Ren He*+ and Juwon Song and Thomas R. Belin
• Companies: UCLA and Korea University and UCLA Department of Biostatistics
• Keywords: Factor analysis ; reversible jump MCMC ; missing data ; Bayesian

Abstract

Factor analysis has been one of the most powerful and flexible tools for assessment of multivariate dependence and codependence. Choosing an appropriate number of factors is very crucial when we conduct factor analysis. when data includes missing values and the data dimenison is high, to choose the proper number of factors is challenging. We propose a reversible jump MCMC method to generate suitable empirical proposal distributions and that address the challenging problem of finding efficient proposals in high-dimensional models with missing data. The simulation study shows that the proposed method select the correct number of factors for simulated data with known number of factors.