Refining exoplanet detection using supervised learning and feature engineering

This work develops a supervised machine learning approach for automating and refining exoplanet detection from astronomical light curves. The method addresses the growing volume of observational data by using feature engineering to derive informative representations of light curves for classification. Rather than replacing expert analysis entirely, the approach is designed to refine the results of case-by-case inspection and automatically classify previously unclassified candidates. Multiple performance criteria are used to evaluate and select suitable features and models. The results demonstrate that supervised learning can effectively generalize across light curves and substantially reduce the time and effort required for manual analysis, supporting scalable automated exoplanet detection.

Recommended citation: Bugueno, Margarita, Francisco Mena, and Mauricio Araya. "Refining exoplanet detection using supervised learning and feature engineering." 2018 XLIV Latin American Computer Conference (CLEI). IEEE, 2018.
Download Slides