Autoencoders for Next-Track-Recommendation
In music recommender systems, playlist continuation is the task of --- given a playlist of a user --- detecting the next track within this playlist (often also referred to as next-track or sequential recommendation). This thesis investigates the suitability and applicability of autoencoders for the task of playlist continuation. We utilize autoencoders and hence, representation learning to complete playlists. Particularly, we design different autoencoders and deep autoencoders for this specific task. Also, we utilize the LFM-1b dataset to evaluate the performance of the proposed approach.