Variable Star Classification Using Multi-View Metric Learning

TL;DR

This paper introduces a multi-view metric learning framework for classifying variable stars, leveraging multi-faceted feature spaces and matrix-variate representations to improve discrimination accuracy.

Contribution

It presents a novel multi-view metric learning approach that extends to matrix-variate data, enhancing variable star classification without prior feature combination.

Findings

Both vector and matrix-variate models perform well on datasets

The framework effectively discriminates variable star categories

Matrix-variate extension offers novel signature representations

Abstract

Our multi-view metric learning framework enables robust characterization of star categories by directly learning to discriminate in a multi-faceted feature space, thus, eliminating the need to combine feature representations prior to fitting the machine learning model. We also demonstrate how to extend standard multi-view learning, which employs multiple vectorized views, to the matrix-variate case which allows very novel variable star signature representations. The performance of our proposed methods is evaluated on the UCR Starlight and LINEAR datasets. Both the vector and matrix-variate versions of our multi-view learning framework perform favorably --- demonstrating the ability to discriminate variable star categories.