Novel Models and Ensemble Techniques to Discriminate Favorite Items from Unrated Ones for Personalized Music Recommendation

Todd G. McKenzie, Chun-Sung Ferng, Yao-Nan Chen, Chun-Liang Li, Cheng-Hao Tsai, Kuan-Wei Wu, Ya-Hsuan Chang, Chung-Yi Li, Wei-Shih Lin, Shu-Hao Yu, Chieh-Yen Lin, Po-Wei Wang, Chia-Mau Ni, Wei-Lun Su, Tsung-Ting Kuo, Chen-Tse Tsai, Po-Lung Chen, Rong-Bing Chiu, Ku-Chun Chou, Yu-Cheng Chou, Chien-Chih Wang, Chen-Hung Wu, Hsuan-Tien Lin, Chih-Jen Lin, Shou-De Lin
; Proceedings of KDD Cup 2011, PMLR 18:101-135, 2012.

Abstract

The Track 2 problem in KDD-Cup 2011 (music recommendation) is to discriminate between music tracks highly rated by a given user from those which are overall highly rated, but not rated by the given user. The training dataset consists of not only user rating history, but also the taxonomic information of track, artist, album, and genre. This paper describes the solution of the National Taiwan University team which ranked first place in the competition. We exploited a diverse of models (neighborhood models, latent models, Bayesian Personalized Ranking models, and random-walk models) with local blending and global ensemble to achieve 97.45% in accuracy on the testing dataset.

Cite this Paper


BibTeX
@InProceedings{pmlr-v18-mckenzie12a, title = {Novel Models and Ensemble Techniques toDiscriminate Favorite Items from Unrated Onesfor Personalized Music Recommendation}, author = {Todd G. McKenzie and Chun-Sung Ferng and Yao-Nan Chen and Chun-Liang Li and Cheng-Hao Tsai and Kuan-Wei Wu and Ya-Hsuan Chang and Chung-Yi Li and Wei-Shih Lin and Shu-Hao Yu and Chieh-Yen Lin and Po-Wei Wang and Chia-Mau Ni and Wei-Lun Su and Tsung-Ting Kuo and Chen-Tse Tsai and Po-Lung Chen and Rong-Bing Chiu and Ku-Chun Chou and Yu-Cheng Chou and Chien-Chih Wang and Chen-Hung Wu and Hsuan-Tien Lin and Chih-Jen Lin and Shou-De Lin}, booktitle = {Proceedings of KDD Cup 2011}, pages = {101--135}, year = {2012}, editor = {Gideon Dror and Yehuda Koren and Markus Weimer}, volume = {18}, series = {Proceedings of Machine Learning Research}, month = {21 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v18/mckenzie12a/mckenzie12a.pdf}, url = {http://proceedings.mlr.press/v18/mckenzie12a.html}, abstract = {The Track 2 problem in KDD-Cup 2011 (music recommendation) is to discriminate between music tracks highly rated by a given user from those which are overall highly rated, but not rated by the given user. The training dataset consists of not only user rating history, but also the taxonomic information of track, artist, album, and genre. This paper describes the solution of the National Taiwan University team which ranked first place in the competition. We exploited a diverse of models (neighborhood models, latent models, Bayesian Personalized Ranking models, and random-walk models) with local blending and global ensemble to achieve 97.45% in accuracy on the testing dataset.} }
Endnote
%0 Conference Paper %T Novel Models and Ensemble Techniques to Discriminate Favorite Items from Unrated Ones for Personalized Music Recommendation %A Todd G. McKenzie %A Chun-Sung Ferng %A Yao-Nan Chen %A Chun-Liang Li %A Cheng-Hao Tsai %A Kuan-Wei Wu %A Ya-Hsuan Chang %A Chung-Yi Li %A Wei-Shih Lin %A Shu-Hao Yu %A Chieh-Yen Lin %A Po-Wei Wang %A Chia-Mau Ni %A Wei-Lun Su %A Tsung-Ting Kuo %A Chen-Tse Tsai %A Po-Lung Chen %A Rong-Bing Chiu %A Ku-Chun Chou %A Yu-Cheng Chou %A Chien-Chih Wang %A Chen-Hung Wu %A Hsuan-Tien Lin %A Chih-Jen Lin %A Shou-De Lin %B Proceedings of KDD Cup 2011 %C Proceedings of Machine Learning Research %D 2012 %E Gideon Dror %E Yehuda Koren %E Markus Weimer %F pmlr-v18-mckenzie12a %I PMLR %J Proceedings of Machine Learning Research %P 101--135 %U http://proceedings.mlr.press %V 18 %W PMLR %X The Track 2 problem in KDD-Cup 2011 (music recommendation) is to discriminate between music tracks highly rated by a given user from those which are overall highly rated, but not rated by the given user. The training dataset consists of not only user rating history, but also the taxonomic information of track, artist, album, and genre. This paper describes the solution of the National Taiwan University team which ranked first place in the competition. We exploited a diverse of models (neighborhood models, latent models, Bayesian Personalized Ranking models, and random-walk models) with local blending and global ensemble to achieve 97.45% in accuracy on the testing dataset.
RIS
TY - CPAPER TI - Novel Models and Ensemble Techniques to Discriminate Favorite Items from Unrated Ones for Personalized Music Recommendation AU - Todd G. McKenzie AU - Chun-Sung Ferng AU - Yao-Nan Chen AU - Chun-Liang Li AU - Cheng-Hao Tsai AU - Kuan-Wei Wu AU - Ya-Hsuan Chang AU - Chung-Yi Li AU - Wei-Shih Lin AU - Shu-Hao Yu AU - Chieh-Yen Lin AU - Po-Wei Wang AU - Chia-Mau Ni AU - Wei-Lun Su AU - Tsung-Ting Kuo AU - Chen-Tse Tsai AU - Po-Lung Chen AU - Rong-Bing Chiu AU - Ku-Chun Chou AU - Yu-Cheng Chou AU - Chien-Chih Wang AU - Chen-Hung Wu AU - Hsuan-Tien Lin AU - Chih-Jen Lin AU - Shou-De Lin BT - Proceedings of KDD Cup 2011 PY - 2012/06/01 DA - 2012/06/01 ED - Gideon Dror ED - Yehuda Koren ED - Markus Weimer ID - pmlr-v18-mckenzie12a PB - PMLR SP - 101 DP - PMLR EP - 135 L1 - http://proceedings.mlr.press/v18/mckenzie12a/mckenzie12a.pdf UR - http://proceedings.mlr.press/v18/mckenzie12a.html AB - The Track 2 problem in KDD-Cup 2011 (music recommendation) is to discriminate between music tracks highly rated by a given user from those which are overall highly rated, but not rated by the given user. The training dataset consists of not only user rating history, but also the taxonomic information of track, artist, album, and genre. This paper describes the solution of the National Taiwan University team which ranked first place in the competition. We exploited a diverse of models (neighborhood models, latent models, Bayesian Personalized Ranking models, and random-walk models) with local blending and global ensemble to achieve 97.45% in accuracy on the testing dataset. ER -
APA
McKenzie, T.G., Ferng, C., Chen, Y., Li, C., Tsai, C., Wu, K., Chang, Y., Li, C., Lin, W., Yu, S., Lin, C., Wang, P., Ni, C., Su, W., Kuo, T., Tsai, C., Chen, P., Chiu, R., Chou, K., Chou, Y., Wang, C., Wu, C., Lin, H., Lin, C. & Lin, S.. (2012). Novel Models and Ensemble Techniques to Discriminate Favorite Items from Unrated Ones for Personalized Music Recommendation. Proceedings of KDD Cup 2011, in PMLR 18:101-135

Related Material