An adaptive algorithm for detecting double stars in astrometric surveys

TL;DR

This paper presents an adaptive machine learning-based method for detecting double stars in astrometric surveys, demonstrating high reliability on HIPPARCOS and Pan-STARRS data, and discusses future platform development.

Contribution

It introduces a novel ML-driven approach for binary star detection using astrometric catalogs, with proven high accuracy and potential for further research platform development.

Findings

Prediction reliability reaches 90-95%

Effective detection demonstrated on HIPPARCOS and Pan-STARRS catalogs

Highlights prospects for creating a dedicated research platform

Abstract

The paper develops a method for detecting optical binary stars based on the use of astrometric catalogs in combination with machine learning (ML) methods. A computational experiment was carried out on the example of the HIPPARCOS mission catalog and the Pan-STARRS (PS1) catalog by applying the suggested method. It has shown that the reliability of predicting a stellar binarity reaches 90-95%. We note the prospects and effectiveness of creating a proprietary research platform - Cognotron.