Citation Context

...re highly correlated. Also, posterior inference tends to be sensitive to certain hyperparameters, with elicitation difficult. To address these issues, Ghosh and Dunson [2009] use parameter expansion [=-=Liu and Wu, 1999-=-, Gelman, 2006] to induce a heavytailed default prior distribution on the loadings elements and propose an efficient Gibbs sampler. Inference on the number of factors in factor analysis models is both...

Citation Context

...ix with non-negative diagonal entries. A popular approach to ensure identifiability of the loadings elements is to constrain the loadings matrix to be lower triangular with positive diagonal entries [=-=Geweke and Zhou, 1996-=-]. Factor models have been traditionally applied in behavioral and social sciences, where the latent factors have a natural interpretation as certain unobserved psychological traits. A more recent app...

Citation Context

... Jordan, 2007]. The Indian buffet process defines a probability distribution on infinite binary matrices, with weighted versions finding applications in factor analysis [Knowles and Ghahramani, 2007, =-=Meeds et al., 2007-=-, Rai and Daumé, 2009]. Our approach can be used as a simpler and more computationally efficient alternative to the Indian buffet process to define priors on latent feature matrices. 2 Bayesian factor...

Citation Context

... Ghahramani, 2006, Thibaux and Jordan, 2007]. The Indian buffet process defines a probability distribution on infinite binary matrices, with weighted versions finding applications in factor analysis [=-=Knowles and Ghahramani, 2007-=-, Meeds et al., 2007, Rai and Daumé, 2009]. Our approach can be used as a simpler and more computationally efficient alternative to the Indian buffet process to define priors on latent feature matrice...

Citation Context

...et al., 2008] uses the above sparse characterization as a dimensionality reduction tool in large p, small n applications such as gene expression studies. A Bayesian specification of the factor model [=-=Arminger and Muthén, 1998-=-, Song and Lee, 2001] commonly uses inverse gamma priors on the residual variances and normal and truncated normal priors on the off-diagonal and diagonal elements of the loadings matrix respectively....

Citation Context

...ee and Song [2002] instead developed a path sampling approach. A more recent method infers the number of factors by zeroing a subset of the loadings elements using Bayesian variable selection priors [=-=Lucas et al., 2006-=-, Carvalho et al., 2008, Schnatter and Lopes, 2009]. Ando [2009] proposed an approach for calculating the exact marginal likelihood in Bayesian factor analysis with heavy-tailed priors. This method ca...

Citation Context

...ting an approximate maximum a posteriori estimate of the covariance matrix. The proposed approach is useful to arrive at a quick working estimate of the covariance matrix. Our proposed Stochastic EM [=-=Celeux et al., 1996-=-] approach replaces draws from the conditional posterior distributions of Λ˜ k (t), Σ and φ in steps 1, 2 and 4 above by the respective conditional posterior modes. 4 Simulation Example 4.1 Factor sel...

Citation Context

...s and propose an efficient Gibbs sampler. Inference on the number of factors in factor analysis models is both conceptually and computationally challenging. Some of the early works in this direction [=-=Polasek, 1997-=-] involve computation of the marginal likelihoods under models with different numbers of factors. Lopes and West [2004] proposed a reversible jump Markov chain Monte Carlo algorithm to allow for uncer...

Citation Context

... sampling approach. A more recent method infers the number of factors by zeroing a subset of the loadings elements using Bayesian variable selection priors [Lucas et al., 2006, Carvalho et al., 2008, =-=Schnatter and Lopes, 2009-=-]. Ando [2009] proposed an approach for calculating the exact marginal likelihood in Bayesian factor analysis with heavy-tailed priors. This method can be used for rapid estimation of the number of fa...