Session-based Music Embedding for Context Aware Recommendation
Nowadays, music streaming platforms are a very popular way to consume music. Due to their popularity, such platforms can be used as a source of information about the listening behavior of their user, which includes the playing history.
The goal of this thesis is to develop a context-aware recommender system that exploits the playing history of a user as context. Therefore, a dataset containing the playing histories of users is split into parts, which represents the playing sequences of each user. The sequences for a single user are divided by their estimated real playing sessions. In a second step, the historical playing sequences of the sessions are used to learn the embeddings of songs. The embedding is used for context aware recommending tracks to users.