Data Analysis of the XBox-2 Radiofrequency Cavity at CERN using Machine Learning Techniques

This master's thesis starts with an introduction to particle physics. Thereby, the basic operating principles of high-gradient linear accelerators are explained. One of the main limitations in these devices is the occurrence of breakdowns, which is investigated in the experimental accelerating structure XBox-2 located at CERN. An adaptable framework for data analysis using machine learning is created with the goal of deriving analysis results from raw experimental data. A strong focus lies on its optimized implementation, which is described in detail. The framework is applied to the data of the XBox-2 accelerator with unsupervised and supervised machine learning techniques. A hypothesis for breakdown indicators is derived from the trained models and tested in the lab. However, further testing on accelerating structures is required before the results of the analysis can be validated.