Matrix Factorization based Heuristics for Constraint-based Recommenders

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Abstract

The main challenges for recommender systems are: producing high quality recommendations and performing many real-time recommendations per second for millions of customers and products. This paper addresses both challenges in the context of constraint-based recommenders where users specify their requirements and the system recommends a solution. We propose a novel approach to determine value ordering heuristics on the basis of matrix factorization. As far as we are aware, no researches exist in constraint-based recommendation domain which exploit matrix factorization techniques. The main idea of our approach consists in the prediction of value ordering heuristics based on historical transactions which can either represent past customer purchases or requirements. Thereby, value ordering heuristics are computed which are specific to each user's requirements. A series of experiments on real-world datasets for calculating constraint-based recommendations has shown that our approach outperforms compared methods in terms of runtime efficiency and prediction quality.
Original languageEnglish
Title of host publicationSAC '19, Proceedings Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing
PublisherAssociation of Computing Machinery
Pages1655-1662
Number of pages8
ISBN (Electronic)978-1-4503-5933-7
DOIs
Publication statusPublished - 2019

Fingerprint

Factorization
Recommender systems
Experiments

Cite this

Polat Erdeniz, S., Felfernig, A., Atas, M., & Samer, R. (2019). Matrix Factorization based Heuristics for Constraint-based Recommenders. In SAC '19, Proceedings Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing (pp. 1655-1662). Association of Computing Machinery. https://doi.org/10.1145/3297280.3297441

Matrix Factorization based Heuristics for Constraint-based Recommenders. / Polat Erdeniz, Seda; Felfernig, Alexander; Atas, Müslüm; Samer, Ralph.

SAC '19, Proceedings Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing. Association of Computing Machinery, 2019. p. 1655-1662.

Research output: Chapter in Book/Report/Conference proceedingConference contributionResearchpeer-review

Polat Erdeniz, S, Felfernig, A, Atas, M & Samer, R 2019, Matrix Factorization based Heuristics for Constraint-based Recommenders. in SAC '19, Proceedings Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing. Association of Computing Machinery, pp. 1655-1662. https://doi.org/10.1145/3297280.3297441
Polat Erdeniz S, Felfernig A, Atas M, Samer R. Matrix Factorization based Heuristics for Constraint-based Recommenders. In SAC '19, Proceedings Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing. Association of Computing Machinery. 2019. p. 1655-1662 https://doi.org/10.1145/3297280.3297441
Polat Erdeniz, Seda ; Felfernig, Alexander ; Atas, Müslüm ; Samer, Ralph. / Matrix Factorization based Heuristics for Constraint-based Recommenders. SAC '19, Proceedings Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing. Association of Computing Machinery, 2019. pp. 1655-1662
@inproceedings{d6e564b08c9545cfb31ab0a957af151e,
title = "Matrix Factorization based Heuristics for Constraint-based Recommenders",
abstract = "The main challenges for recommender systems are: producing high quality recommendations and performing many real-time recommendations per second for millions of customers and products. This paper addresses both challenges in the context of constraint-based recommenders where users specify their requirements and the system recommends a solution. We propose a novel approach to determine value ordering heuristics on the basis of matrix factorization. As far as we are aware, no researches exist in constraint-based recommendation domain which exploit matrix factorization techniques. The main idea of our approach consists in the prediction of value ordering heuristics based on historical transactions which can either represent past customer purchases or requirements. Thereby, value ordering heuristics are computed which are specific to each user's requirements. A series of experiments on real-world datasets for calculating constraint-based recommendations has shown that our approach outperforms compared methods in terms of runtime efficiency and prediction quality.",
author = "{Polat Erdeniz}, Seda and Alexander Felfernig and M{\"u}sl{\"u}m Atas and Ralph Samer",
year = "2019",
doi = "10.1145/3297280.3297441",
language = "English",
pages = "1655--1662",
booktitle = "SAC '19, Proceedings Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing",
publisher = "Association of Computing Machinery",
address = "United States",

}

TY - GEN

T1 - Matrix Factorization based Heuristics for Constraint-based Recommenders

AU - Polat Erdeniz, Seda

AU - Felfernig, Alexander

AU - Atas, Müslüm

AU - Samer, Ralph

PY - 2019

Y1 - 2019

N2 - The main challenges for recommender systems are: producing high quality recommendations and performing many real-time recommendations per second for millions of customers and products. This paper addresses both challenges in the context of constraint-based recommenders where users specify their requirements and the system recommends a solution. We propose a novel approach to determine value ordering heuristics on the basis of matrix factorization. As far as we are aware, no researches exist in constraint-based recommendation domain which exploit matrix factorization techniques. The main idea of our approach consists in the prediction of value ordering heuristics based on historical transactions which can either represent past customer purchases or requirements. Thereby, value ordering heuristics are computed which are specific to each user's requirements. A series of experiments on real-world datasets for calculating constraint-based recommendations has shown that our approach outperforms compared methods in terms of runtime efficiency and prediction quality.

AB - The main challenges for recommender systems are: producing high quality recommendations and performing many real-time recommendations per second for millions of customers and products. This paper addresses both challenges in the context of constraint-based recommenders where users specify their requirements and the system recommends a solution. We propose a novel approach to determine value ordering heuristics on the basis of matrix factorization. As far as we are aware, no researches exist in constraint-based recommendation domain which exploit matrix factorization techniques. The main idea of our approach consists in the prediction of value ordering heuristics based on historical transactions which can either represent past customer purchases or requirements. Thereby, value ordering heuristics are computed which are specific to each user's requirements. A series of experiments on real-world datasets for calculating constraint-based recommendations has shown that our approach outperforms compared methods in terms of runtime efficiency and prediction quality.

U2 - 10.1145/3297280.3297441

DO - 10.1145/3297280.3297441

M3 - Conference contribution

SP - 1655

EP - 1662

BT - SAC '19, Proceedings Proceedings of the 34th ACM/SIGAPP Symposium on Applied Computing

PB - Association of Computing Machinery

ER -