Exploration and evaluation of different sessionization methods in a music streaming context

Abstract: Sessionization is a powerful method to create aggregated data used to understanduser behavior in the field of data mining. A lot of research has been performed overthe years and most of it is geared towards traditional web servers. How the resultstranslate to a context with a stand alone application with a service that has veryspecific usage pattern has not been fully investigated. In this project we investigatedhow the currently utilized heuristics perform on identifying a session in a musicstreaming context and introduce some additional measures to summarize such sessions.