A Linear Ensemble of Individual and Blended Models for Music Rating Prediction

Po-Lung Chen, Chen-Tse Tsai, Yao-Nan Chen, Ku-Chun Chou, Chun-Liang Li, Cheng-Hao Tsai, Kuan-Wei Wu, Yu-Cheng Chou, Chung-Yi Li, Wei-Shih Lin, Shu-Hao Yu, Rong-Bing Chiu, Chieh-Yen Lin, Chien-Chih Wang, Po-Wei Wang, Wei-Lun Su, Chen-Hung Wu, Tsung-Ting Kuo, Todd G. McKenzie, Ya-Hsuan Chang, Chun-Sung Ferng, Chia-Mau Ni, Hsuan-Tien Lin, Chih-Jen Lin, Shou-De Lin
Proceedings of KDD Cup 2011, PMLR 18:21-60, 2012.

Abstract

Track 1 of KDDCup 2011 aims at predicting the rating behavior of users in the Yahoo! Music system. At National Taiwan University, we organize a course that teams up students to work on both tracks of KDDCup 2011. For track 1, we first tackle the problem by building variants of existing individual models, including Matrix Factorization, Restricted Boltzmann Machine, k-Nearest Neighbors, Probabilistic Latent Semantic Analysis, Probabilistic Principle Component Analysis and Supervised Regression. We then blend the individual models along with some carefully extracted features in a non-linear manner. A large linear ensemble that contains both the individual and the blended models is learned and taken through some post-processing steps to form the final solution. The four stages: individual model building, non-linear blending, linear ensemble and post-processing lead to a successful final solution, within which techniques on feature engineering and aggregation (blending and ensemble learning) play crucial roles. Our team is the first prize winner of both tracks of KDD Cup 2011.

Cite this Paper


BibTeX
@InProceedings{pmlr-v18-chen12a, title = {A Linear Ensemble of Individual and Blended Models for Music Rating Prediction}, author = {Chen, Po-Lung and Tsai, Chen-Tse and Chen, Yao-Nan and Chou, Ku-Chun and Li, Chun-Liang and Tsai, Cheng-Hao and Wu, Kuan-Wei and Chou, Yu-Cheng and Li, Chung-Yi and Lin, Wei-Shih and Yu, Shu-Hao and Chiu, Rong-Bing and Lin, Chieh-Yen and Wang, Chien-Chih and Wang, Po-Wei and Su, Wei-Lun and Wu, Chen-Hung and Kuo, Tsung-Ting and McKenzie, Todd G. and Chang, Ya-Hsuan and Ferng, Chun-Sung and Ni, Chia-Mau and Lin, Hsuan-Tien and Lin, Chih-Jen and Lin, Shou-De}, booktitle = {Proceedings of KDD Cup 2011}, pages = {21--60}, year = {2012}, editor = {Dror, Gideon and Koren, Yehuda and Weimer, Markus}, volume = {18}, series = {Proceedings of Machine Learning Research}, month = {21 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v18/chen12a/chen12a.pdf}, url = {https://proceedings.mlr.press/v18/chen12a.html}, abstract = {Track 1 of KDDCup 2011 aims at predicting the rating behavior of users in the Yahoo! Music system. At National Taiwan University, we organize a course that teams up students to work on both tracks of KDDCup 2011. For track 1, we first tackle the problem by building variants of existing individual models, including Matrix Factorization, Restricted Boltzmann Machine, k-Nearest Neighbors, Probabilistic Latent Semantic Analysis, Probabilistic Principle Component Analysis and Supervised Regression. We then blend the individual models along with some carefully extracted features in a non-linear manner. A large linear ensemble that contains both the individual and the blended models is learned and taken through some post-processing steps to form the final solution. The four stages: individual model building, non-linear blending, linear ensemble and post-processing lead to a successful final solution, within which techniques on feature engineering and aggregation (blending and ensemble learning) play crucial roles. Our team is the first prize winner of both tracks of KDD Cup 2011.} }
Endnote
%0 Conference Paper %T A Linear Ensemble of Individual and Blended Models for Music Rating Prediction %A Po-Lung Chen %A Chen-Tse Tsai %A Yao-Nan Chen %A Ku-Chun Chou %A Chun-Liang Li %A Cheng-Hao Tsai %A Kuan-Wei Wu %A Yu-Cheng Chou %A Chung-Yi Li %A Wei-Shih Lin %A Shu-Hao Yu %A Rong-Bing Chiu %A Chieh-Yen Lin %A Chien-Chih Wang %A Po-Wei Wang %A Wei-Lun Su %A Chen-Hung Wu %A Tsung-Ting Kuo %A Todd G. McKenzie %A Ya-Hsuan Chang %A Chun-Sung Ferng %A Chia-Mau Ni %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-chen12a %I PMLR %P 21--60 %U https://proceedings.mlr.press/v18/chen12a.html %V 18 %X Track 1 of KDDCup 2011 aims at predicting the rating behavior of users in the Yahoo! Music system. At National Taiwan University, we organize a course that teams up students to work on both tracks of KDDCup 2011. For track 1, we first tackle the problem by building variants of existing individual models, including Matrix Factorization, Restricted Boltzmann Machine, k-Nearest Neighbors, Probabilistic Latent Semantic Analysis, Probabilistic Principle Component Analysis and Supervised Regression. We then blend the individual models along with some carefully extracted features in a non-linear manner. A large linear ensemble that contains both the individual and the blended models is learned and taken through some post-processing steps to form the final solution. The four stages: individual model building, non-linear blending, linear ensemble and post-processing lead to a successful final solution, within which techniques on feature engineering and aggregation (blending and ensemble learning) play crucial roles. Our team is the first prize winner of both tracks of KDD Cup 2011.
RIS
TY - CPAPER TI - A Linear Ensemble of Individual and Blended Models for Music Rating Prediction AU - Po-Lung Chen AU - Chen-Tse Tsai AU - Yao-Nan Chen AU - Ku-Chun Chou AU - Chun-Liang Li AU - Cheng-Hao Tsai AU - Kuan-Wei Wu AU - Yu-Cheng Chou AU - Chung-Yi Li AU - Wei-Shih Lin AU - Shu-Hao Yu AU - Rong-Bing Chiu AU - Chieh-Yen Lin AU - Chien-Chih Wang AU - Po-Wei Wang AU - Wei-Lun Su AU - Chen-Hung Wu AU - Tsung-Ting Kuo AU - Todd G. McKenzie AU - Ya-Hsuan Chang AU - Chun-Sung Ferng AU - Chia-Mau Ni AU - Hsuan-Tien Lin AU - Chih-Jen Lin AU - Shou-De Lin BT - Proceedings of KDD Cup 2011 DA - 2012/06/01 ED - Gideon Dror ED - Yehuda Koren ED - Markus Weimer ID - pmlr-v18-chen12a PB - PMLR DP - Proceedings of Machine Learning Research VL - 18 SP - 21 EP - 60 L1 - http://proceedings.mlr.press/v18/chen12a/chen12a.pdf UR - https://proceedings.mlr.press/v18/chen12a.html AB - Track 1 of KDDCup 2011 aims at predicting the rating behavior of users in the Yahoo! Music system. At National Taiwan University, we organize a course that teams up students to work on both tracks of KDDCup 2011. For track 1, we first tackle the problem by building variants of existing individual models, including Matrix Factorization, Restricted Boltzmann Machine, k-Nearest Neighbors, Probabilistic Latent Semantic Analysis, Probabilistic Principle Component Analysis and Supervised Regression. We then blend the individual models along with some carefully extracted features in a non-linear manner. A large linear ensemble that contains both the individual and the blended models is learned and taken through some post-processing steps to form the final solution. The four stages: individual model building, non-linear blending, linear ensemble and post-processing lead to a successful final solution, within which techniques on feature engineering and aggregation (blending and ensemble learning) play crucial roles. Our team is the first prize winner of both tracks of KDD Cup 2011. ER -
APA
Chen, P., Tsai, C., Chen, Y., Chou, K., Li, C., Tsai, C., Wu, K., Chou, Y., Li, C., Lin, W., Yu, S., Chiu, R., Lin, C., Wang, C., Wang, P., Su, W., Wu, C., Kuo, T., McKenzie, T.G., Chang, Y., Ferng, C., Ni, C., Lin, H., Lin, C. & Lin, S.. (2012). A Linear Ensemble of Individual and Blended Models for Music Rating Prediction. Proceedings of KDD Cup 2011, in Proceedings of Machine Learning Research 18:21-60 Available from https://proceedings.mlr.press/v18/chen12a.html.

Related Material