Gamma and Hadron Particle Classification Using Physics Feature Extraction and Optuna Hyperparameter Tuning.

When using a ground-based Imaging Atmospheric Cherenkov Telescope like MAGIC to capture or detect high-energy gamma rays greater than 100 GeVs, the hadronic noise and background activities from protons and heavy nuclei pose a challenge in gamma photonic energy detection. Thus, to address this difficulty in VHE astronomy in distinguishing between gamma rays and hadron rays within the electromagnetic shower of cosmic rays, we propose a novel articulated approach. Herein, we shall consider a lightweight optimization through ’Optuna’ hyperparameter tuning and using the XGBoost ensemble technique, post performing physics-feature extraction based on IACT protocols and guidelines for MAGIC Telescopic Data Analysis. By leveraging the power of this approach, we were able to achieve an accuracy of 89.33 %.