TY - GEN
T1 - Exploiting Performance Estimates for Augmenting Recommendation Ensembles
AU - Penha, Gustavo
AU - Santos, Rodrygo L.T.
PY - 2020
Y1 - 2020
N2 - Ensembling multiple recommender systems via stacking has shown to be effective at improving collaborative recommendation. Recent work extends stacking to use additional user performance predictors (e.g., the total number of ratings made by the user) to help determine how much each base recommender should contribute to the ensemble. Nonetheless, despite the cost of handcrafting discriminative predictors, which typically requires deep knowledge of the strengths and weaknesses of each recommender in the ensemble, only minor improvements have been observed. To overcome this limitation, instead of engineering complex features to predict the performance of different recommenders for a given user, we propose to directly estimate these performances by leveraging the user's own historical ratings. Experiments on real-world datasets from multiple domains demonstrate that using performance estimates as additional features can significantly improve the accuracy of state-of-the-art ensemblers, achieving nDCG@20 improvements by an average of 23% over not using them.
AB - Ensembling multiple recommender systems via stacking has shown to be effective at improving collaborative recommendation. Recent work extends stacking to use additional user performance predictors (e.g., the total number of ratings made by the user) to help determine how much each base recommender should contribute to the ensemble. Nonetheless, despite the cost of handcrafting discriminative predictors, which typically requires deep knowledge of the strengths and weaknesses of each recommender in the ensemble, only minor improvements have been observed. To overcome this limitation, instead of engineering complex features to predict the performance of different recommenders for a given user, we propose to directly estimate these performances by leveraging the user's own historical ratings. Experiments on real-world datasets from multiple domains demonstrate that using performance estimates as additional features can significantly improve the accuracy of state-of-the-art ensemblers, achieving nDCG@20 improvements by an average of 23% over not using them.
KW - Ensembling
KW - Learning to Rank
KW - Performance Estimation
KW - Performance Prediction
KW - Recommender Systems
UR - http://www.scopus.com/inward/record.url?scp=85092745801&partnerID=8YFLogxK
U2 - 10.1145/3383313.3412264
DO - 10.1145/3383313.3412264
M3 - Conference contribution
AN - SCOPUS:85092745801
T3 - RecSys 2020 - 14th ACM Conference on Recommender Systems
SP - 111
EP - 119
BT - RecSys 2020 - 14th ACM Conference on Recommender Systems
PB - Association for Computing Machinery (ACM)
T2 - 14th ACM Conference on Recommender Systems, RecSys 2020
Y2 - 22 September 2020 through 26 September 2020
ER -