Generalized Canonical Correlation Analysis for Classification

TL;DR

This paper introduces conditions under which Generalized Canonical Correlation Analysis (GCCA) enhances classification accuracy when integrating multiple datasets, supported by simulations and real data experiments.

Contribution

It provides theoretical conditions for GCCA's superiority over standard CCA in multi-dataset classification tasks.

Findings

GCCA can outperform CCA in classification accuracy under certain conditions.

Theoretical analysis of GCCA's benefits for multi-view data integration.

Empirical validation through simulations and real data experiments.

Abstract

For multiple multivariate data sets, we derive conditions under which Generalized Canonical Correlation Analysis (GCCA) improves classification performance of the projected datasets, compared to standard Canonical Correlation Analysis (CCA) using only two data sets. We illustrate our theoretical results with simulations and a real data experiment.