Higher Order Correlation Analysis for Multi-View Learning

TL;DR

This paper introduces a novel higher order correlation maximization approach for multi-view learning, addressing the limitations of pairwise correlation methods by capturing intrinsic interconnections among multiple views.

Contribution

It formulates higher order correlation as a low rank tensor approximation problem and employs a generating polynomial method to solve it, improving multi-view learning performance.

Findings

Outperforms existing methods on real multi-view datasets

Effectively captures intrinsic interconnections among multiple views

Demonstrates consistent improvement in correlation maximization

Abstract

Multi-view learning is frequently used in data science. The pairwise correlation maximization is a classical approach for exploring the consensus of multiple views. Since the pairwise correlation is inherent for two views, the extensions to more views can be diversified and the intrinsic interconnections among views are generally lost. To address this issue, we propose to maximize higher order correlations. This can be formulated as a low rank approximation problem with the higher order correlation tensor of multi-view data. We use the generating polynomial method to solve the low rank approximation problem. Numerical results on real multi-view data demonstrate that this method consistently outperforms prior existing methods.