[Back]

J. Martin Gonzalvez:
"Detection of User Profiles and Tariff Classes in Mobile Networks Using Machine Learning Algorithms";
Supervisor: M. Rupp, P. Svoboda, C. Midoglu; Institute of Telecommunications, TU Wien, 2016; final examination: 06-01-2016.

The aim of the Master Thesis is to employ machine learning techniques such as feature analysis and clustering on an open data set of crowdsourced measurements, for purposes of mobile network benchmarking. In particular, different user profiles in mobile networks will be investigated through these methods. With the purpose of mapping tariff classes to detected user profiles, a performance related list of features will be presented and the best clustering algorithm will be established by comparing results from all those which have been selected.

The aim of the Master Thesis is to employ machine learning techniques such as feature analysis and clustering on an open data set of crowdsourced measurements, for purposes of mobile network benchmarking. In particular, different user profiles in mobile networks will be investigated through these methods. With the purpose of mapping tariff classes to detected user profiles, a performance related list of features will be presented and the best clustering algorithm will be established by comparing results from all those which have been selected.

clustering, feature analysis, crowdsourcing, Expectation-Maximization, K-Means, tariff classes