Gaussian processes for canonical correlation analysis

We consider several stochastic process methods for performing canonical correlation analysis (CCA). The first uses a Gaussian Process formulation of regression in which we use the current projection of one data set as the target for the other and then repeat in the opposite direction. The second uses a method which relies on probabilistically sphering the data, concatenating the two streams and then performing a probabilistic PCA. The third gets the canonical correlation projections directly without having to calculate the filters first. We also investigate nonlinearity and sparsification of these methods. Gaussian Processes for Canonical Correlation Analysis Colin Fyfe, Gayle Leen and Pei Ling Lai 1. Applied Computational Intelligence Research Unit, The University of Paisley, Scotland. email:colin.fyfe,[email protected] 2. Southern Taiwan University of Technology, Tainan, Taiwan email:pei ling [email protected] Abstract We consider several stochastic process methods for performing canonical correlation analysis (CCA). The first uses a Gaussian Process formulation of regression in which we use the current projection of one data set as the target for the other and then repeat in the opposite direction. The second uses a method which relies on probabilistically sphering the data, concatenating the two streams and then performing a probabilistic PCA. The third gets the canonical correlation projections directly without having to calculate the filters first. We also investigate nonlinearity and sparsification of these methods.We consider several stochastic process methods for performing canonical correlation analysis (CCA). The first uses a Gaussian Process formulation of regression in which we use the current projection of one data set as the target for the other and then repeat in the opposite direction. The second uses a method which relies on probabilistically sphering the data, concatenating the two streams and then performing a probabilistic PCA. The third gets the canonical correlation projections directly without having to calculate the filters first. We also investigate nonlinearity and sparsification of these methods.