Music recommendation via heterogeneous information graph embedding

Publication Type:
Conference Proceeding
Proceedings of the International Joint Conference on Neural Networks, 2017, 2017-May pp. 596 - 603
Issue Date:
Filename Description Size
07965907 (002).pdfAccepted Manuscript382.61 kB
Adobe PDF
Full metadata record
© 2017 IEEE. Traditional music recommendation techniques suffer from limited performance due to the sparsity of user-music interaction data, which is addressed by incorporating auxiliary information. In this paper, we study the problem of personalized music recommendation that takes different kinds of auxiliary information into consideration. To achieve this goal, a Heterogeneous Information Graph (HIG) is first constructed to encode different kinds of heterogeneous information, including the interactions between users and music pieces, music playing sequences, and the metadata of music pieces. Based on HIG, a Heterogeneous Information Graph Embedding method (HIGE) is proposed to learn the latent low-dimensional representations of music pieces. Then, we further develop a context-aware music recommendation method. Extensive experiments have been conducted on real-world datasets to compare the proposed method with other state-of-the-art recommendation methods. The results demonstrate that the proposed method significantly outperforms those baselines, especially on sparse datasets.
Please use this identifier to cite or link to this item: