The thesis deals with algorithms for detecting anomalies in the data collected by the Problem Solving Tutor research. In the theoretical part, the author introduces the term of anomaly, the ideas of the PROSO research and a detailed overview of existing algorithms to detect anomalies. In the practical part, selected algorithms are implemented. Real outputs and results as well as recommendations to a user are presented in this part, and the chapter is supplemented by a number of graphs. The implemented algorithms are also compared to existing data mining software. An example of working with the data mining tools, applied to the data coming from PROSO, and explanation of their outputs is also part of the thesis. In the summary, the appropriate methodology for behavior analysis and anomaly detection is determinied.
Identifer | oai:union.ndltd.org:nusl.cz/oai:invenio.nusl.cz:178666 |
Date | January 2014 |
Creators | Šormová, Hana |
Source Sets | Czech ETDs |
Language | Czech |
Detected Language | English |
Type | info:eu-repo/semantics/masterThesis |
Rights | info:eu-repo/semantics/restrictedAccess |
Page generated in 0.3306 seconds