Abstract
Music therapy for improving recognition ability may be more effective when the favorite music of each person is adopted. In the proposed system, first, the recommendation process using collaborative filtering is terminated when no users in the reference list have the same preference of recommended music as that of a new user. Then, the second recommendation process finds the most similar music, from the scores for impression words, to those successfully recommended among music not recommended up to the moment. The average number of recommended songs for each user by the proposed system was 12.1, whereas that of collaborative filtering was 4.3. The recommendation accuracy of the proposed system was 70.2 %, whereas that of collaborative filtering was 62.1 %. The ratings of songs can be added on a user-by-user basis in the recommendation process, and this increased number of cases improves the recommendation accuracy and increases the number of recommended songs.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
References
Konishi H, Yoshitomi Y (2011) Music recommendation system aimed at improving recognition ability. In: Sugisaka M (ed), Proceedings of 16th International Symposium on Artificial Life and Robotics (AROB16th), Beppu, Oita Japan, 27–29, Jan, 2011, pp 241–244
Koro C, Yoshitomi Y, Asada T, Yoshizaki S (2012), Music recommendation aimed at improving recognition ability using collaborative filtering and impression words. In: Sugisaka M (ed), Proceedings of 17th International Symposium on Artificial Life and Robotics (AROB17th), Beppu, Oita Japan, 19–21, Jan 2012, pp 222–225
Kumamoto T, Ohta K (2002) Design of scales to represent user’s impressions of a music piece for use in a music-retrieval system (in Japanese). IPSJ SIG Notes, 2001-NL-147(6):35–40
Takahashi T (1997) Research report on songs familiar to people advanced in years (in Japanese). J Jpn Music Ther Assoc 15(1):68–75
Akahoshi T (2009) Good old anthology enjoyable for people advanced in years and troubled with dementia (in Japanese). Kirara shobo, Tokyo
Music textbook database for elementary school by Kanagawa prefectural education center (in Japanese). http://kjd.edu-ctr.pref.kanagawa.jp/daizai_music/ Accessed 11 March 2013
Acknowledgments
We would like to thank Associate Professor M. Tabuse of Kyoto Prefectural University for his valuable advice and support. We would also like to thank all the participants who cooperated with us in the experiments. This work was partially supported by SCOPE (122307003) of Ministry of Internal Affairs of Communications of Japan government.
Author information
Authors and Affiliations
Corresponding author
About this article
Cite this article
Yoshizaki, S., Yoshitomi, Y., Koro, C. et al. Music recommendation hybrid system for improving recognition ability using collaborative filtering and impression words. Artif Life Robotics 18, 109–116 (2013). https://doi.org/10.1007/s10015-013-0107-z
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10015-013-0107-z