This paper presents a content-based approach to music recommendation. For this purpose, a database which contains more than 50000 music excerpts acquired from public repositories was built. Datasets contain tracks of distinct performers within several music genres. All music pieces were converted to mp3 format and then parameterized based on MPEG-7, mel-cepstral and time-related dedicated parameters. All feature vectors are stored as csv files and will be available on-line. A study of the database statistical characteristics was performed. Different splits into train and test sets were investigated to provide the most accurate evaluation of the decision-based solutions. Classification time and memory complexity were also evaluated.
Author:
Kostek, Bozena
Affiliation:
Gdansk University of Technology, Gdansk, Poland
AES Convention:
131 (October 2011)
Paper Number:
8505
Publication Date:
October 19, 2011
Subject:
Posters: Production and Analysis of Musical Sounds
Click to purchase paper as a non-member or you can login as an AES member to see more options.
No AES members have commented on this paper yet.
To be notified of new comments on this paper you can subscribe to this RSS feed. Forum users should login to see additional options.
If you are not yet an AES member and have something important to say about this paper then we urge you to join the AES today and make your voice heard. You can join online today by clicking here.