A Bayesian sparse factor model with adaptive posterior concentration
This work addresses a foundational challenge in statistical modeling for high-dimensional data, offering an incremental improvement in Bayesian inference for sparse factor models.
The paper tackles the problem of inferring factor dimensionality and sparse structure in high-dimensional factor models by proposing a Bayesian method that introduces dependence between sparsity and dimensionality, achieving adaptive posterior concentration. It shows asymptotic posterior consistency, optimal detection rates, and near-optimal covariance concentration, with numerical studies demonstrating superiority over competitors.
In this paper, we propose a new Bayesian inference method for a high-dimensional sparse factor model that allows both the factor dimensionality and the sparse structure of the loading matrix to be inferred. The novelty is to introduce a certain dependence between the sparsity level and the factor dimensionality, which leads to adaptive posterior concentration while keeping computational tractability. We show that the posterior distribution asymptotically concentrates on the true factor dimensionality, and more importantly, this posterior consistency is adaptive to the sparsity level of the true loading matrix and the noise variance. We also prove that the proposed Bayesian model attains the optimal detection rate of the factor dimensionality in a more general situation than those found in the literature. Moreover, we obtain a near-optimal posterior concentration rate of the covariance matrix. Numerical studies are conducted and show the superiority of the proposed method compared with other competitors.