Classification of blazars based on data-driven approaches
Active galactic nuclei (AGNs), including blazars, exhibit distinctive variability in their optical light curves, making them ideal for classification studies. This work uses data from the latest GAIA and Pan-STARRS data releases to analyze these patterns. The goal of this work is to classify AGNs into two categories: "blazars" and "non-blazars'' using only optical light curves. This strategy differs from most existing works, as it relies exclusively on optical variability without employing any other multiwavelength information. We processed optical light curves from GAIA and Pan-STARRS using the FATS library to extract standard time-series features. We computed additional features with custom algorithms based on literature methods. A Light Gradient-Boosting Machine (LightGBM) model was trained to classify AGNs into blazars and non-blazars based on these features. We then used this knowledge base to carry out a self-learning experiment with AGN candidates of an unknown nature. The LightGBM model achieved an accuracy of $86\%$, with precision, recall, and F1 score above $80-85\%$ for classifying blazars and non-blazar AGNs using optical data. The application of a BoostBoruta algorithm for feature selection reduced the feature space from 70 to 13. while maintaining comparable performance. A self-training classifier yielded similar results $85\%$, confirming the robustness of the model and the reliability of pseudo-labeling for unknown objects.