Abstract: A music recommendation system was developed that can learn users' preferences. The system can classify a wide range of stored music using automatic music content analyses. Users can opt for music according to their mood, using such words as "bright", "exciting", "quiet", and "sad". Building a music recommendation system is one of the information retrieval tasks. This research is devoted to a content-based music recommender system. The main peculiarity of our work is that the developed recommender system is based on the acoustic similarity of musical compositions. Two approaches to building a content-based music recommender system are considered in this paper. The first is a quite common approach that uses acoustic features analysis. The second approach includes deep learning and computer vision methods applications aimed at improving the results of the recommender system.
Keywords: Numpy, Pandas, Cosine Similarity, Count Vectorizer
|
DOI:
10.17148/IJARCCE.2023.124115
