Collaborative Filtering Ensemble for Ranking

Michael Jahrer, Andreas Töscher
Proceedings of KDD Cup 2011, PMLR 18:153-167, 2012.

Abstract

This paper provides the solution of the team “commendo” on the Track2 dataset of the KDD Cup 2011 Dror et al.. Yahoo Labs provides a snapshot of their music-rating database as dataset for the competition, consisting of approximately 62 million ratings from 250k users on 300k items. The dataset includes hierachical information about the items. The goal of the competition is to distinguish beteen “High rated” and “Not rated” items of a user. The rating scale is discrete and ranges from 0 to 100, while a “High” rating is a rating$\geq 0$. The error measure is the percent of false rated tracks over all users, known as the fractions of misclassifications. The task is to minimize this error rate, hence the ranking should be optimized. Our final submission is a blend of different collaborative filtering algorithms enhanced, with basic statistics. The algorithms are trained consecutively and they are blended together with a neural network. Each of the algorithms optimizes a rank error measure.

Cite this Paper

BibTeX
@InProceedings{pmlr-v18-jahrer12b, title = {Collaborative Filtering Ensemble for Ranking}, author = {Jahrer, Michael and Töscher, Andreas}, booktitle = {Proceedings of KDD Cup 2011}, pages = {153--167}, 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/jahrer12b/jahrer12b.pdf}, url = {https://proceedings.mlr.press/v18/jahrer12b.html}, abstract = {This paper provides the solution of the team “commendo” on the Track2 dataset of the KDD Cup 2011 Dror et al.. Yahoo Labs provides a snapshot of their music-rating database as dataset for the competition, consisting of approximately 62 million ratings from 250k users on 300k items. The dataset includes hierachical information about the items. The goal of the competition is to distinguish beteen “High rated” and “Not rated” items of a user. The rating scale is discrete and ranges from 0 to 100, while a “High” rating is a rating$\geq 0$. The error measure is the percent of false rated tracks over all users, known as the fractions of misclassifications. The task is to minimize this error rate, hence the ranking should be optimized. Our final submission is a blend of different collaborative filtering algorithms enhanced, with basic statistics. The algorithms are trained consecutively and they are blended together with a neural network. Each of the algorithms optimizes a rank error measure.} }
Endnote
%0 Conference Paper %T Collaborative Filtering Ensemble for Ranking %A Michael Jahrer %A Andreas Töscher %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-jahrer12b %I PMLR %P 153--167 %U https://proceedings.mlr.press/v18/jahrer12b.html %V 18 %X This paper provides the solution of the team “commendo” on the Track2 dataset of the KDD Cup 2011 Dror et al.. Yahoo Labs provides a snapshot of their music-rating database as dataset for the competition, consisting of approximately 62 million ratings from 250k users on 300k items. The dataset includes hierachical information about the items. The goal of the competition is to distinguish beteen “High rated” and “Not rated” items of a user. The rating scale is discrete and ranges from 0 to 100, while a “High” rating is a rating$\geq 0$. The error measure is the percent of false rated tracks over all users, known as the fractions of misclassifications. The task is to minimize this error rate, hence the ranking should be optimized. Our final submission is a blend of different collaborative filtering algorithms enhanced, with basic statistics. The algorithms are trained consecutively and they are blended together with a neural network. Each of the algorithms optimizes a rank error measure.
RIS
TY - CPAPER TI - Collaborative Filtering Ensemble for Ranking AU - Michael Jahrer AU - Andreas Töscher BT - Proceedings of KDD Cup 2011 DA - 2012/06/01 ED - Gideon Dror ED - Yehuda Koren ED - Markus Weimer ID - pmlr-v18-jahrer12b PB - PMLR DP - Proceedings of Machine Learning Research VL - 18 SP - 153 EP - 167 L1 - http://proceedings.mlr.press/v18/jahrer12b/jahrer12b.pdf UR - https://proceedings.mlr.press/v18/jahrer12b.html AB - This paper provides the solution of the team “commendo” on the Track2 dataset of the KDD Cup 2011 Dror et al.. Yahoo Labs provides a snapshot of their music-rating database as dataset for the competition, consisting of approximately 62 million ratings from 250k users on 300k items. The dataset includes hierachical information about the items. The goal of the competition is to distinguish beteen “High rated” and “Not rated” items of a user. The rating scale is discrete and ranges from 0 to 100, while a “High” rating is a rating$\geq 0$. The error measure is the percent of false rated tracks over all users, known as the fractions of misclassifications. The task is to minimize this error rate, hence the ranking should be optimized. Our final submission is a blend of different collaborative filtering algorithms enhanced, with basic statistics. The algorithms are trained consecutively and they are blended together with a neural network. Each of the algorithms optimizes a rank error measure. ER -
APA
Jahrer, M. & Töscher, A.. (2012). Collaborative Filtering Ensemble for Ranking. Proceedings of KDD Cup 2011, in Proceedings of Machine Learning Research 18:153-167 Available from https://proceedings.mlr.press/v18/jahrer12b.html.

Related Material