Multilinear Class-Specific Discriminant Analysis

TL;DR

This paper introduces a novel multilinear subspace learning method that enhances class-specific discrimination in tensor data, preserving spatial structure and improving performance in facial analysis and stock prediction.

Contribution

It proposes a class-specific multilinear discriminant analysis technique for tensor data, filling a gap in existing methods that focus on inter- and intra-class variance.

Findings

Effective in facial image analysis

Improves stock price prediction accuracy

Preserves spatial structure of tensor data

Abstract

There has been a great effort to transfer linear discriminant techniques that operate on vector data to high-order data, generally referred to as Multilinear Discriminant Analysis (MDA) techniques. Many existing works focus on maximizing the inter-class variances to intra-class variances defined on tensor data representations. However, there has not been any attempt to employ class-specific discrimination criteria for the tensor data. In this paper, we propose a multilinear subspace learning technique suitable for applications requiring class-specific tensor models. The method maximizes the discrimination of each individual class in the feature space while retains the spatial structure of the input. We evaluate the efficiency of the proposed method on two problems, i.e. facial image analysis and stock price prediction based on limit order book data.