Variational Bounds for Mixed-Data Factor Analysis
Emtiyaz Khan, Guillaume Bouchard, Ben Marlin, Kevin Murphy
We propose a new variational EM algorithm for fitting factor analysis models
with mixed continuous and categorical observations. The algorithm is based on a
simple quadratic bound to the log-sum-exp function. In the special case of fully observed binary data, the bound we propose is significantly faster than previous variational methods. We show that EM is significantly more robust in the presence of missing data compared to treating the latent factors as parameters, which is the approach used by exponential family PCA and other related matrix-factorization methods. A further benefit of the variational approach is that it can easily be extended to the case of mixtures of factor analyzers, as we show. We present results on synthetic and real data sets demonstrating several desirable properties of our proposed method.
NIPS (Neural Information Processing Systems), Vancouver, B.C., Canada, December 6th, 2010 Full paper available on NIPS 2010 Proceedings