Automated classification of Hipparcos unsolved variables

TL;DR

This paper develops an automated system using machine learning to classify various types of variable stars in the Hipparcos catalogue, including irregular and non-periodic variables, achieving high accuracy.

Contribution

It introduces a multistage classification approach combining random forests and Bayesian networks, expanding the classification to 24 variability types with improved accuracy.

Findings

Misclassification rates under 12% for the classifiers.

Successful prediction of variability types for 6051 Hipparcos variables.

Enhanced classification of irregular and non-periodic variables.

Abstract

We present an automated classification of stars exhibiting periodic, non-periodic and irregular light variations. The Hipparcos catalogue of unsolved variables is employed to complement the training set of periodic variables of Dubath et al. with irregular and non-periodic representatives, leading to 3881 sources in total which describe 24 variability types. The attributes employed to characterize light-curve features are selected according to their relevance for classification. Classifier models are produced with random forests and a multistage methodology based on Bayesian networks, achieving overall misclassification rates under 12 per cent. Both classifiers are applied to predict variability types for 6051 Hipparcos variables associated with uncertain or missing types in the literature.