Home // IMMM 2012, The Second International Conference on Advances in Information Mining and Management // View article

Music Recommendation Based on Text Mining

Authors:
Ziwon Hyung
MyoungA Lee
Kyogu Lee

Keywords: text mining; Latent Semantic Analysis; music recommendation

Abstract:
Recommending music from millions of items is a challenging problem. In this paper, we propose a novel approach to recommending music given an textual input from the user. To this end, we first mine a large corpus of textual documents from the radio station's Internet bulletin board. Each document, written by a listener, contains a personal story associated with a song request. Assuming that the personal story contains the reason for the song request, we then perform the Latent Semantic Analysis (LSA) on these documents to find the document similarity, which we believe also indicates similar music preference. Our hypothesis is that when the two users request the same song, the situation or context in which they write the associated story is likely to be similar as well, and therefore the two stories will also be similar to each other. Using the mined documents that request the same song as a test set, we show that there is a positive correlation between the document similarity and song similarity, and thus it is possible to recommend music purely based on text mining and analysis.

Pages: 129 to 134

Copyright: Copyright (c) IARIA, 2012

Publication date: October 21, 2012

Published in: conference

ISSN: 2326-9332

ISBN: 978-1-61208-227-1

Location: Venice, Italy

Dates: from October 21, 2012 to October 26, 2012