Unsupervised Time Series Outlier Detection

Student thesis: Master programme thesis

  • David Gonzalo Chaves Campos
4. term, Computer Science (IT), Master (Master Programme)
The continuous advance in processes digitalization has led to an extended availability of sensor-based devices that aims to improve and complete tasks efficiently in many domains. The breakthrough of using digital methods is driven by the massive production of data in form of time series, which allows steady monitoring of operations to take action when some relevant condition is reached, for instance, anomalies or outliers. Achieving that needs to develop automatic methods with no supervision since the available data is vast. The problem is addressed in this work, considering robust state-of-the-art methodologies and infrastructure with extensive experiments using real-world data sets from different domains.
LanguageEnglish
Publication date4 Jun 2021
ID: 413847667