Semi-supervised Multi-sensor Classification via Consensus-based Multi-View Maximum Entropy Discrimination

TL;DR

This paper introduces CMV-MED, a semi-supervised multi-sensor classification method that leverages consensus among classifiers to improve accuracy using unlabeled data within a multi-view learning framework.

Contribution

It proposes a novel consensus-based multi-view maximum entropy discrimination algorithm for semi-supervised multi-sensor classification.

Findings

Improved classification accuracy over previous methods.

Effective utilization of unlabeled data in multi-sensor settings.

Validated on three real multi-sensor datasets.

Abstract

In this paper, we consider multi-sensor classification when there is a large number of unlabeled samples. The problem is formulated under the multi-view learning framework and a Consensus-based Multi-View Maximum Entropy Discrimination (CMV-MED) algorithm is proposed. By iteratively maximizing the stochastic agreement between multiple classifiers on the unlabeled dataset, the algorithm simultaneously learns multiple high accuracy classifiers. We demonstrate that our proposed method can yield improved performance over previous multi-view learning approaches by comparing performance on three real multi-sensor data sets.