Master theses
Probabilistic and Machine Learning Models for Anomaly Detection in Time Series Data
Author
Anna Husieva
Year
2024
Type
Master thesis
Supervisor
Ing. Vojtěch Šalanský, Ph.D.
Reviewers
doc. Ing. Kamil Dedecius, Ph.D.
Department
Summary
This thesis examines the effectiveness of probabilistic and machine learning techniques in anomaly detection within time series data. A review of relevant literature lays the groundwork for implementing and evaluating two probabilistic and three machine learning models. Through rigorous experimentation on both real-world and synthetic datasets, the strengths and weaknesses of these models are identified across various anomaly types. A hybrid approach is proposed and implemented, integrating elements from both probabilistic and machine learning frameworks. Its efficacy in enhancing anomaly detection performance is assessed.