Approximate modeling of continuous context in factorization algorithms

Balázs Hidasi, D. Tikk

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

3 Citations (Scopus)

Abstract

Factorization based algorithms - such as matrix or tensor factorization - are widely used in the field of recommender systems. These methods model the relations between the entities of two or more dimensions. The entity based approach is suitable for dimensions such as users, items and several context types, where the domain of the context is nominal. Continuous and ordinal context dimensions are usually discretized and their values are used as nominal entities. While this enables the usage of continuous context in factorization methods, still much information is lost during the process. In this paper we propose two approaches for better modeling of the continuous context dimensions. Fuzzy event modeling tackles the problem through the uncertainty of the value of the observation in the context dimension. Fuzzy context modeling, on the other hand, enables context-states to overlap, thus certain observations are influenced by multiple context-states. Throughout the paper seasonality is used as an example of continuous context. We incorporate the modeling concepts into the iTALS algorithm, without degrading its scalability. The effect of the two approaches on recommendation accuracy is measured on five implicit feedback databases.

Original languageEnglish
Title of host publicationACM International Conference Proceeding Series
PublisherAssociation for Computing Machinery
Pages3-9
Number of pages7
ISBN (Print)9781450327237
DOIs
Publication statusPublished - 2014
Event4th Workshop on Context-Awareness in Retrieval and Recommendation, CaRR 2014 - In Conjunction with the European Conference on Information Retrieval, ECIR 2014 - Amsterdam, Netherlands
Duration: Apr 13 2014Apr 13 2014

Other

Other4th Workshop on Context-Awareness in Retrieval and Recommendation, CaRR 2014 - In Conjunction with the European Conference on Information Retrieval, ECIR 2014
CountryNetherlands
CityAmsterdam
Period4/13/144/13/14

Fingerprint

Factorization
Recommender systems
Tensors
Scalability
Feedback

Keywords

  • context-awareness
  • continuous context
  • factorization
  • recommender systems

ASJC Scopus subject areas

  • Human-Computer Interaction
  • Computer Networks and Communications
  • Computer Vision and Pattern Recognition
  • Software

Cite this

Hidasi, B., & Tikk, D. (2014). Approximate modeling of continuous context in factorization algorithms. In ACM International Conference Proceeding Series (pp. 3-9). Association for Computing Machinery. https://doi.org/10.1145/2601301.2601303

Approximate modeling of continuous context in factorization algorithms. / Hidasi, Balázs; Tikk, D.

ACM International Conference Proceeding Series. Association for Computing Machinery, 2014. p. 3-9.

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

Hidasi, B & Tikk, D 2014, Approximate modeling of continuous context in factorization algorithms. in ACM International Conference Proceeding Series. Association for Computing Machinery, pp. 3-9, 4th Workshop on Context-Awareness in Retrieval and Recommendation, CaRR 2014 - In Conjunction with the European Conference on Information Retrieval, ECIR 2014, Amsterdam, Netherlands, 4/13/14. https://doi.org/10.1145/2601301.2601303
Hidasi B, Tikk D. Approximate modeling of continuous context in factorization algorithms. In ACM International Conference Proceeding Series. Association for Computing Machinery. 2014. p. 3-9 https://doi.org/10.1145/2601301.2601303
Hidasi, Balázs ; Tikk, D. / Approximate modeling of continuous context in factorization algorithms. ACM International Conference Proceeding Series. Association for Computing Machinery, 2014. pp. 3-9
@inproceedings{5de9ab1d887443cba05825cfca8817f7,
title = "Approximate modeling of continuous context in factorization algorithms",
abstract = "Factorization based algorithms - such as matrix or tensor factorization - are widely used in the field of recommender systems. These methods model the relations between the entities of two or more dimensions. The entity based approach is suitable for dimensions such as users, items and several context types, where the domain of the context is nominal. Continuous and ordinal context dimensions are usually discretized and their values are used as nominal entities. While this enables the usage of continuous context in factorization methods, still much information is lost during the process. In this paper we propose two approaches for better modeling of the continuous context dimensions. Fuzzy event modeling tackles the problem through the uncertainty of the value of the observation in the context dimension. Fuzzy context modeling, on the other hand, enables context-states to overlap, thus certain observations are influenced by multiple context-states. Throughout the paper seasonality is used as an example of continuous context. We incorporate the modeling concepts into the iTALS algorithm, without degrading its scalability. The effect of the two approaches on recommendation accuracy is measured on five implicit feedback databases.",
keywords = "context-awareness, continuous context, factorization, recommender systems",
author = "Bal{\'a}zs Hidasi and D. Tikk",
year = "2014",
doi = "10.1145/2601301.2601303",
language = "English",
isbn = "9781450327237",
pages = "3--9",
booktitle = "ACM International Conference Proceeding Series",
publisher = "Association for Computing Machinery",

}

TY - GEN

T1 - Approximate modeling of continuous context in factorization algorithms

AU - Hidasi, Balázs

AU - Tikk, D.

PY - 2014

Y1 - 2014

N2 - Factorization based algorithms - such as matrix or tensor factorization - are widely used in the field of recommender systems. These methods model the relations between the entities of two or more dimensions. The entity based approach is suitable for dimensions such as users, items and several context types, where the domain of the context is nominal. Continuous and ordinal context dimensions are usually discretized and their values are used as nominal entities. While this enables the usage of continuous context in factorization methods, still much information is lost during the process. In this paper we propose two approaches for better modeling of the continuous context dimensions. Fuzzy event modeling tackles the problem through the uncertainty of the value of the observation in the context dimension. Fuzzy context modeling, on the other hand, enables context-states to overlap, thus certain observations are influenced by multiple context-states. Throughout the paper seasonality is used as an example of continuous context. We incorporate the modeling concepts into the iTALS algorithm, without degrading its scalability. The effect of the two approaches on recommendation accuracy is measured on five implicit feedback databases.

AB - Factorization based algorithms - such as matrix or tensor factorization - are widely used in the field of recommender systems. These methods model the relations between the entities of two or more dimensions. The entity based approach is suitable for dimensions such as users, items and several context types, where the domain of the context is nominal. Continuous and ordinal context dimensions are usually discretized and their values are used as nominal entities. While this enables the usage of continuous context in factorization methods, still much information is lost during the process. In this paper we propose two approaches for better modeling of the continuous context dimensions. Fuzzy event modeling tackles the problem through the uncertainty of the value of the observation in the context dimension. Fuzzy context modeling, on the other hand, enables context-states to overlap, thus certain observations are influenced by multiple context-states. Throughout the paper seasonality is used as an example of continuous context. We incorporate the modeling concepts into the iTALS algorithm, without degrading its scalability. The effect of the two approaches on recommendation accuracy is measured on five implicit feedback databases.

KW - context-awareness

KW - continuous context

KW - factorization

KW - recommender systems

UR - http://www.scopus.com/inward/record.url?scp=84903600697&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84903600697&partnerID=8YFLogxK

U2 - 10.1145/2601301.2601303

DO - 10.1145/2601301.2601303

M3 - Conference contribution

AN - SCOPUS:84903600697

SN - 9781450327237

SP - 3

EP - 9

BT - ACM International Conference Proceeding Series

PB - Association for Computing Machinery

ER -