This work is part of a collaboration between the Music Techology Group and Kakao Corp, under the supervision of Xavier Serra.
Team: cocoplaya
Members:
- Andres Ferraro (MTG)
- Dmitry Bogdanov (MTG)
- Jason Yoon (Kakao Corp)
- Lucas Kim (Kakao Corp)
Username in recsys-challenge.spotify.com: aferraro
contact: andres.ferraro at upf.edu
In order to reproduce the recommendations submitted for the Creative track of the challenge the following steps must be executed:
- Step 1: Download audio samples of the songs from spotify
- Step 2: Compute features from the audio of the songs
- Step 3: Train Matrix Factorization model and generate recommendations, merge with reccomendations generated from songs co-ocurrence probabilities.
The code to reproduce each step is contained in a separate folder, you can find further instructions inside the folders.
For citations you can reference this paper::
Ferraro, A., Bogdanov, D., Yoon, J., Kim, K., & Serra, X. (2018, October). "Automatic playlist continuation using a hybrid recommender system combining features from text and audio". In Proceedings of the ACM Recommender Systems Challenge 2018 (p. 2). ACM.