Music Information Retrieval Recommender System Audio Features
Issue Date:
22-Jun-2017
Publisher:
Institute of Mathematics and Informatics Bulgarian Academy of Sciences, Association for the Development of the Information Society
Citation:
Proceedings of the National Conference on "Education and Research in the Information Society", Plovdiv, June, 2017, 031p-037p
Series/Report no.:
ADIS;2017
Abstract:
The paper presents a research that aims to investigate whether sound features can
be used for recommending music. First it presents a study of existing tools for sound processing
in order to see what features of the sound can be extracted with these tools. Second it presents
experiments that use machine learning algorithms to identify the key features of the sound for
the purpose of recommending music. Finally, manually classified data from 19 users were used
for experiments. The achieved maximum average accuracy was measured to be 68.16%. This
is an 18.17% increase in accuracy over the baseline. The conclusion is that it makes sense to
analyze sound for the purpose of recommending music.
Description:
Report published in the Proceedings of the National Conference on "Education and Research in the Information Society", Plovdiv, June, 2017