ABSTRACT
Automated playlist generation is a special form of music recommendation and a common feature of digital music playing applications. A particular challenge of the task is that the recommended items should not only match the general listener's preference but should also be coherent with the most recently played tracks. In this work, we propose a novel algorithmic approach and optimization scheme to generate playlist continuations that address these requirements. In our approach, we first use collections of shared music playlists, music metadata, and user preferences to select suitable tracks with high accuracy. Next, we apply a generic re-ranking optimization scheme to generate playlist continuations that match the characteristics of the last played tracks. An empirical evaluation on three collections of shared playlists shows that the combination of different input signals helps to achieve high accuracy during track selection and that the re-ranking technique can both help to balance different quality optimization goals and to further increase accuracy.
- G. Adomavicius and Y. Kwon. Improving aggregate recommendation diversity using ranking-based techniques. IEEE TKDE, 24(5):896--911, 2012. Google ScholarDigital Library
- W. Balkema and F. van der Heijden. Music Playlist Generation by Assimilating GMMs into SOMs. Pattern Recognition Letters, 31(11):1396--1402, 2010. Google ScholarDigital Library
- G. Bonnin and D. Jannach. Evaluating the Quality of Playlists Based on Hand-Crafted Samples. In Proc. ISMIR, pages 263--268, 2013.Google Scholar
- G. Bonnin and D. Jannach. Automated generation of music playlists: Survey and experiments. ACM Comput. Surv., 47(2):26:1--26:35, 2014. Google ScholarDigital Library
- K. Bradley and B. Smyth. Improving Recommendation Diversity. In Proc. AICS '01, pages 75--84, 2001.Google Scholar
- P. Y. K. Chau, S. Y. Ho, K. K. W. Ho, and Y. Yao. Examining the effects of malfunctioning personalized services on online users' distrust and behaviors. Decis. Support Syst., 56:180--191, 2013. Google ScholarDigital Library
- S. Cunningham, D. Bainbridge, and A. Falconer. 'More of an Art than a Science': Supporting the Creation of Playlists and Mixes. In Proc. ISMIR, pages 240--245, 2006.Google Scholar
- N. Hariri, B. Mobasher, and R. Burke. Context-Aware Music Recommendation Based on Latent Topic Sequential Patterns. In Proc. RecSys, pages 131--138, 2012. Google ScholarDigital Library
- T. Jambor and J. Wang. Optimizing multiple objectives in collaborative filtering. In Proc. RecSys '10, pages 55--62, 2010. Google ScholarDigital Library
- D. Jannach, I. Kamehkhosh, and G. Bonnin. Analyzing the characteristics of shared playlists for music recommendation. In RSWeb Workhop at ACM RecSys '14, 2014.Google Scholar
- D. Jannach, L. Lerche, F. Gedikli, and G. Bonnin. What recommenders recommend - an analysis of accuracy, popularity, and sales diversity effects. In Proc. UMAP 2013, pages 25--37, 2013.Google ScholarCross Ref
- M. Kaminskas and F. Ricci. Contextual music information retrieval and recommendation: State of the art and challenges. Computer Science Review, 6(2--3):89--119, 2012.Google ScholarCross Ref
- B. Logan. Content-Based Playlist Generation: Exploratory Experiments. In Proc. ISMIR, pages 295--296, 2002.Google Scholar
- B. McFee and G. R. Lanckriet. The Natural Language of Playlists. In Proc. ISMIR, pages 537--542, 2011.Google Scholar
- B. McFee and G. R. Lanckriet. Hypergraph Models of Playlist Dialects. In Proc. ISMIR, pages 343--348, 2012.Google Scholar
- S. Pauws, W. Verhaegh, and M. Vossen. Music playlist generation by adapted simulated annealing. Inf. Sci., 178(3):647--662, 2008. Google ScholarDigital Library
- S. Rendle, C. Freudenthaler, Z. Gantner, and L. Schmidt-Thieme. BPR: Bayesian personalized ranking from implicit feedback. In Proc. UAI '09, pages 452--461, 2009. Google ScholarDigital Library
- A. Said, B. J. Jain, and S. Albayrak. A 3D approach to recommender system evaluation. In Proc. CSCW '13 Companion Volume, pages 263--266, 2013. Google ScholarDigital Library
- A. M. Sarroff and M. Casey. Modeling and Predicting Song Adjacencies In Commercial Albums. In Proc. SMC, 2012.Google Scholar
- M. Zhang and N. Hurley. Avoiding monotony: Improving the diversity of recommendation lists. In RecSys '08, pages 123--130, 2008. Google ScholarDigital Library
- Y. C. Zhang, D. O. Séaghdha, D. Quercia, and T. Jambor. Auralist: Introducing serendipity into music recommendation. In Proc. WSDM '12, pages 13--22, 2012. Google ScholarDigital Library
- C.-N. Ziegler, S. M. McNee, J. A. Konstan, and G. Lausen. Improving recommendation lists through topic diversification. In Proc. WWW '05, pages 22--32, 2005. Google ScholarDigital Library
Index Terms
- Beyond "Hitting the Hits": Generating Coherent Music Playlist Continuations with the Right Tracks
Recommendations
Biases in Automated Music Playlist Generation: A Comparison of Next-Track Recommending Techniques
UMAP '16: Proceedings of the 2016 Conference on User Modeling Adaptation and PersonalizationPlaylist generation is a special form of music recommendation where the problem is to create a sequence of tracks to be played next, given a number of seed tracks. In academia, the evaluation of playlisting techniques is often done by assessing with the ...
Tailoring Music Recommendations to Users by Considering Diversity, Mainstreaminess, and Novelty
SIGIR '15: Proceedings of the 38th International ACM SIGIR Conference on Research and Development in Information RetrievalA shortcoming of current approaches for music recommendation is that they consider user-specific characteristics only on a very simple level, typically as some kind of interaction between users and items when employing collaborative filtering. To ...
Leveraging multi-dimensional user models for personalized next-track music recommendation
SAC '17: Proceedings of the Symposium on Applied ComputingThe automated generation of playlists given a user's last played tracks is a common feature of modern music platforms. Existing approaches to this "next-track music recommendation" problem often focus solely on the user's recent listening behavior or ...
Comments