Vector projection on lyrics and user comments for a lightweight emotion-aware Chinese music recommendation system

Document Type

Conference Proceeding

Publication Date



With the development of modern internet, information explodes exponentially, which makes people hard to find out what information they really want in a huge data set. In many famous music platforms in China, such as QQ Music, Wangyi Cloud Music, Xiami Music, how to recommend appropriate music to users is a very challenging issue. Most of these recommendation systems are using collaborative filtering method to recommend music. In our paper, we propose a light-weight emotion-aware approach which can analyze emotion based on the song lyric and user comments, via vector projection to decide which song can be recommended to the users in order to maximize their experiences. Two emotion lexicons had been adopted for such purposes. To validate our system, a small-scale user study has been conducted with mixed and interesting results. We offer suggestions on the adoption of each lexicon in the context of user satisfaction and use.

Publication Title

ACM International Conference Proceeding Series

First Page Number


Last Page Number




This document is currently not available here.