Sparsification of Probabilistic Canonical Correlation Analysis

We have recently developed several ways of performing Canonical Correlation Analysis [1, 5, 7, 4] with probabilistic methods rather than the standard statistical tools. However, the computational demands of training such methods scales with the square of the number of samples, making these methods uncompetitive with e.g. artificial neural network methods [3, 2]. In this paper, we examine a recent development which sparsifies a probabilistic method of performing principal component analysis and then use this method to sparsify a new probabilistic method of performing canonical correlation analysis. Key-words: Probabilistic canonical correlation analysis, sparsification