Assignment of master’s thesis - Probabilistic and Machine Learning Models for Anomaly Detection in Time Series Data

Aug 3, 2026

Upload by Miles

This master’s thesis studies the effectiveness of probabilistic techniques and machine learning methods for anomaly detection in time series data. A structured review of relevant literature supports the implementation and evaluation of multiple models, including probabilistic and machine learning approaches. Experiments are conducted on both real-world and synthetic datasets to identify strengths and limitations across different anomaly scenarios. A hybrid method integrating probabilistic and machine learning elements is proposed, implemented, and assessed for its ability to improve detection performance.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages87

Reading Time3 h 39 min

AccessFree to download

Instructions

Literature review and model selection

Model implementation and evaluation

Comparative strengths and weaknesses

Hybrid probabilistic–machine learning approach and comparison

Title: Assignment of master’s thesis Probabilistic and Machine Learning Models for Anomaly Detection in Time Series Data Student: Bc. Anna Husieva Supervisor: Ing. Vojtěch Šalanský, Ph. D. Study program: Informatics Branch / specialization: Knowledge Engineering Department: Department of Applied Mathematics Validity: until the end of summer semester 2024/2025 Instructions 1) Review and compile relevant research articles and papers on both probabilistic and machine learning methods for anomaly detection in time series data such as [1, 2] . 2) Implement at least two probabilistic models and two machine learning models tailored for time series data and evaluate their performance on several public available datasets such as [3, 4, 5] . 3) Investigate the strengths and weaknesses of choosen models across diverse anomaly scenarios. 4) Propose and implement a method that combines probabilistic and machine learning models. Determine its effectiveness in enhancing anomaly detection performance across different types of anomalies. Compare the results of the proposed method with the selected models mentioned above. [1] Schmidl, Sebastian, Phillip Wenig and Thorsten Papenbrock.“Anomaly Detection in Time Series: A Comprehensive Evaluation.” Proc. VLDB Endow. 15 (2022): 1779-1797. [2] Braei, Mohammad and Sebastian Wagner.“Anomaly Detection in Univariate Timeseries: A Survey on the State-of-the-Art.” ArXiv abs/2004 .00433 (2020) [3] Lavin, Alexander and Subutai Ahmad.“Evaluating Real-Time Anomaly Detection Algorithms--The Numenta Anomaly Benchmark.” 2015 IEEE 14th International Conference on Machine Learning and Applications (ICMLA) (2015): 38-44. Electronically approved by Ing. Magda Friedjungová, Ph.D. on 22 December 2023 in Prague. [4] Phillip Wenig, Sebastian Schmidl, and Thorsten Papenbrock. TimeEval: A Benchmarking Toolkit for Time Series Anomaly Detection Algorithms. PVLDB, 15(12): 3678 -3681, 2022 [5] A Labeled Anomaly Detection Dataset (ydata-labeled-time-series-anomalies-v1 _ 0), Yahoo! Webscope Program. 2020 [http://labs.yahoo.com/](http://labs.yahoo.com/) Electronically approved by Ing. Magda Friedjungová, Ph.D. on 22 December 2023 in Prague. Master’s thesis Probabilistic and Machine Learning Models for Anomaly Detection in Time Series Data Bc . Anna Husieva Department of Applied Mathematics Supervisor: Ing. Vojtěch Šalanský, Ph.D. May 8, 2024 Acknowledgements First and foremost, I would like to express my heartfelt appreciation to my thesis supervisor, Ing. Vojtěch Šalanský, Ph.D. , whose invaluable guidance and insightful advice, as well as his patience and willingness to help, have significantly contributed to the success of my research. I would also like to extend my deepest gratitude to my family, whose unwavering support has been instrumental in enabling me to have the opportunity to achieve this milestone. A special thanks to my father, whose resilience and determination always inspired me to never give up. Declaration I hereby declare that the presented thesis is my own work and that I have cited all sources of information in accordance with the Guideline for adhering to ethical principles when elaborating an academic final thesis. I acknowledge that my thesis is subject to the rights and obligations stipulated by the Act No. 121/2000 Coll., the Copyright Act, as amended, in particular that the Czech Technical University in Prague has the right to conclude a license agreement on the utilization of this thesis as a school work under the provisions of Article 60 (1) of the Act. In Prague on May 8, 2024 . . .. . .. . .. . .. . .. . .. . . Czech Technical University in Prague Faculty of Information Technology © 2024 Anna Husieva. All rights reserved. This thesis is school work as defined by Copyright Act of the Czech Republic. It has been submitted at Czech Technical University in Prague, Faculty of Information Technology. The thesis is protected by the Copyr

FAQs

What is the main objective of the thesis?

The thesis aims to evaluate probabilistic and machine learning models for anomaly detection in time series data and to improve performance using a hybrid approach.

Which models does the thesis require implementing and evaluating?

It requires implementing at least two probabilistic models and at least two machine learning models tailored to time series, then evaluating their performance on public datasets.

How does the thesis propose to enhance anomaly detection performance?

It proposes and implements a method that combines probabilistic and machine learning models, then compares its results against the selected individual models across different anomaly types.

Aug 3, 2026

Upload by Miles