Theoretical Mathematics & Applications

A Predictive Model for Canonical Correlation Analysis with Implications for the Simple and Multiple Correlations

  • Pdf Icon [ Download ]
  • Times downloaded: 9870
  •                                                    Abstract

    This study presents a new formulation for canonical correlation analysis as an equivalent multiple regression model for two Gaussian random vectors. In addition, implications for the simple and multiple correlations are discussed. The new model can also be extended to measure nonlinear associations between sets of variables.