A unified approach of factor models and neighbor based methods for large recommender systems

Gábor Takács, István Pilászy, Bottyán Németh, D. Tikk

Research output: Chapter in Book/Report/Conference proceedingConference contribution

10 Citations (Scopus)

Abstract

Matrix factorization (MF) based approaches have proven to be efficient for rating-based recommendation systems. In this paper, we propose a hybrid approach that alloys an improved MF and the so-called NSVD1 approach, resulting in a very accurate factor model. After that, we propose a unification of factor models and neighbor based approaches, which further improves the performance. The approaches are evaluated on the Netflix Prize dataset, and they provide very low RMSE, and favorable running time. Our best solution presented here with Quiz RMSE 0.8851 outperforms all published single methods in the literature.

Original languageEnglish
Title of host publication1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008
Pages186-191
Number of pages6
DOIs
Publication statusPublished - 2008
Event1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008 - Ostrava, Czech Republic
Duration: Aug 4 2008Aug 6 2008

Other

Other1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008
CountryCzech Republic
CityOstrava
Period8/4/088/6/08

Fingerprint

Recommender systems
Factorization

ASJC Scopus subject areas

  • Computer Science Applications
  • Information Systems
  • Software

Cite this

Takács, G., Pilászy, I., Németh, B., & Tikk, D. (2008). A unified approach of factor models and neighbor based methods for large recommender systems. In 1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008 (pp. 186-191). [4664342] https://doi.org/10.1109/ICADIWT.2008.4664342

A unified approach of factor models and neighbor based methods for large recommender systems. / Takács, Gábor; Pilászy, István; Németh, Bottyán; Tikk, D.

1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008. 2008. p. 186-191 4664342.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Takács, G, Pilászy, I, Németh, B & Tikk, D 2008, A unified approach of factor models and neighbor based methods for large recommender systems. in 1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008., 4664342, pp. 186-191, 1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008, Ostrava, Czech Republic, 8/4/08. https://doi.org/10.1109/ICADIWT.2008.4664342
Takács G, Pilászy I, Németh B, Tikk D. A unified approach of factor models and neighbor based methods for large recommender systems. In 1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008. 2008. p. 186-191. 4664342 https://doi.org/10.1109/ICADIWT.2008.4664342
Takács, Gábor ; Pilászy, István ; Németh, Bottyán ; Tikk, D. / A unified approach of factor models and neighbor based methods for large recommender systems. 1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008. 2008. pp. 186-191
@inproceedings{66c367712b6c4ffda6b053527165a903,
title = "A unified approach of factor models and neighbor based methods for large recommender systems",
abstract = "Matrix factorization (MF) based approaches have proven to be efficient for rating-based recommendation systems. In this paper, we propose a hybrid approach that alloys an improved MF and the so-called NSVD1 approach, resulting in a very accurate factor model. After that, we propose a unification of factor models and neighbor based approaches, which further improves the performance. The approaches are evaluated on the Netflix Prize dataset, and they provide very low RMSE, and favorable running time. Our best solution presented here with Quiz RMSE 0.8851 outperforms all published single methods in the literature.",
author = "G{\'a}bor Tak{\'a}cs and Istv{\'a}n Pil{\'a}szy and Botty{\'a}n N{\'e}meth and D. Tikk",
year = "2008",
doi = "10.1109/ICADIWT.2008.4664342",
language = "English",
isbn = "9781424426249",
pages = "186--191",
booktitle = "1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008",

}

TY - GEN

T1 - A unified approach of factor models and neighbor based methods for large recommender systems

AU - Takács, Gábor

AU - Pilászy, István

AU - Németh, Bottyán

AU - Tikk, D.

PY - 2008

Y1 - 2008

N2 - Matrix factorization (MF) based approaches have proven to be efficient for rating-based recommendation systems. In this paper, we propose a hybrid approach that alloys an improved MF and the so-called NSVD1 approach, resulting in a very accurate factor model. After that, we propose a unification of factor models and neighbor based approaches, which further improves the performance. The approaches are evaluated on the Netflix Prize dataset, and they provide very low RMSE, and favorable running time. Our best solution presented here with Quiz RMSE 0.8851 outperforms all published single methods in the literature.

AB - Matrix factorization (MF) based approaches have proven to be efficient for rating-based recommendation systems. In this paper, we propose a hybrid approach that alloys an improved MF and the so-called NSVD1 approach, resulting in a very accurate factor model. After that, we propose a unification of factor models and neighbor based approaches, which further improves the performance. The approaches are evaluated on the Netflix Prize dataset, and they provide very low RMSE, and favorable running time. Our best solution presented here with Quiz RMSE 0.8851 outperforms all published single methods in the literature.

UR - http://www.scopus.com/inward/record.url?scp=57949091675&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=57949091675&partnerID=8YFLogxK

U2 - 10.1109/ICADIWT.2008.4664342

DO - 10.1109/ICADIWT.2008.4664342

M3 - Conference contribution

SN - 9781424426249

SP - 186

EP - 191

BT - 1st International Conference on the Applications of Digital Information and Web Technologies, ICADIWT 2008

ER -