A machine learning approach to gamma-ray identification with the Cherenkov Telescope Array

Kaal, Wiegert (2023) A machine learning approach to gamma-ray identification with the Cherenkov Telescope Array. Bachelor's Thesis, Astronomy.

Preview	Text BSc_Thesis_Wiegert_Kaal_s4491904.pdf Download (11MB) | Preview
	Text toestemming.pdf Restricted to Registered users only Download (131kB)

Abstract

The Cherenkov Telescope Array (CTA) is expected to be far more accurate and have a five to ten times higher sensitivity than the current generation of ground-based gamma-ray detectors. In this work, we tested the effectiveness of Boosted Decision Trees (BDTs), a widely used Machine Learning algorithm that is easy to implement and has shown promising results in similar classification problems, for the separation of gamma-ray-initiated air showers and cosmic-ray-initiated air showers, which constitute the main background for gamma-ray astronomy, being about 1000 times more abundant. We looked at what properties of the air shower image were the most important for the BDT, and how accurately it could identify gamma-ray showers for CTA. “Out of a total sample of 39 features which were provided in the simulated events provided by the CTA consortium, we identified 4 features that provided a good separation power. Using all 39 features, our BDT reaches a precision of over 98 percent. This makes the BDT a very suitable method for the separation of gamma-ray-initiated and cosmic-ray-initiated air showers.

Item Type:	Thesis (Bachelor's Thesis)
Supervisor name:	Vecchi, M.
Degree programme:	Astronomy
Thesis type:	Bachelor's Thesis
Language:	English
Date Deposited:	10 Jul 2023 12:07
Last Modified:	10 Jul 2023 12:07
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/30425

Actions (login required)

View Item