Revealing and incorporating implicit communities to improve recommender systems

Euijin Choo, Ting Yu, Min Chi, Yan Sun

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

10 Citations (Scopus)

Abstract

Social connections often have a significant influence on personal decision making. Researchers have proposed novel recommender systems that take advantage of social relationship information to improve recommendations. These systems, while promising, are often hindered in practice. Existing social networks such as Facebook are not designed for recommendations and thus contain many irrelevant relationships. Many recommendation platforms such as Amazon often do not permit users to establish explicit social relationships. And direct integration of social and commercial systems raises privacy concerns. In this paper we address these issues by focusing on the extraction of implicit and relevant relationships among users based upon the patterns of their existing interactions. Our work is grounded in the context of item recommendations on Amazon. We investigate whether users' reply patterns can be used to identify these meaningful relationships and show that different degrees of relationships do exist. We develop global measures of relationship strength and observe that users tend to form strong connections when they are evaluating subjective items such as books and movies. We then design a probabilistic mechanism to distinguish meaningful connections from connections formed by chance and extract implicit communities. We finally show that these communities can be used for hybrid recommender systems that improve recommendations over existing collaborative filtering approaches.

Original languageEnglish
Title of host publicationEC 2014 - Proceedings of the 15th ACM Conference on Economics and Computation
PublisherAssociation for Computing Machinery
Pages489-506
Number of pages18
ISBN (Print)9781450325653
DOIs
Publication statusPublished - 1 Jan 2014
Event15th ACM Conference on Economics and Computation, EC 2014 - Palo Alto, CA, United States
Duration: 8 Jun 201412 Jun 2014

Other

Other15th ACM Conference on Economics and Computation, EC 2014
CountryUnited States
CityPalo Alto, CA
Period8/6/1412/6/14

Fingerprint

Recommender systems
Collaborative filtering
Decision making

Keywords

  • implicit communities
  • recommendation
  • recommender system
  • social network

ASJC Scopus subject areas

  • Computer Science (miscellaneous)

Cite this

Choo, E., Yu, T., Chi, M., & Sun, Y. (2014). Revealing and incorporating implicit communities to improve recommender systems. In EC 2014 - Proceedings of the 15th ACM Conference on Economics and Computation (pp. 489-506). Association for Computing Machinery. https://doi.org/10.1145/2600057.2602906

Revealing and incorporating implicit communities to improve recommender systems. / Choo, Euijin; Yu, Ting; Chi, Min; Sun, Yan.

EC 2014 - Proceedings of the 15th ACM Conference on Economics and Computation. Association for Computing Machinery, 2014. p. 489-506.

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

Choo, E, Yu, T, Chi, M & Sun, Y 2014, Revealing and incorporating implicit communities to improve recommender systems. in EC 2014 - Proceedings of the 15th ACM Conference on Economics and Computation. Association for Computing Machinery, pp. 489-506, 15th ACM Conference on Economics and Computation, EC 2014, Palo Alto, CA, United States, 8/6/14. https://doi.org/10.1145/2600057.2602906
Choo E, Yu T, Chi M, Sun Y. Revealing and incorporating implicit communities to improve recommender systems. In EC 2014 - Proceedings of the 15th ACM Conference on Economics and Computation. Association for Computing Machinery. 2014. p. 489-506 https://doi.org/10.1145/2600057.2602906
Choo, Euijin ; Yu, Ting ; Chi, Min ; Sun, Yan. / Revealing and incorporating implicit communities to improve recommender systems. EC 2014 - Proceedings of the 15th ACM Conference on Economics and Computation. Association for Computing Machinery, 2014. pp. 489-506
@inproceedings{d4c262ba1ec244c38ba84a5e5f3d3f26,
title = "Revealing and incorporating implicit communities to improve recommender systems",
abstract = "Social connections often have a significant influence on personal decision making. Researchers have proposed novel recommender systems that take advantage of social relationship information to improve recommendations. These systems, while promising, are often hindered in practice. Existing social networks such as Facebook are not designed for recommendations and thus contain many irrelevant relationships. Many recommendation platforms such as Amazon often do not permit users to establish explicit social relationships. And direct integration of social and commercial systems raises privacy concerns. In this paper we address these issues by focusing on the extraction of implicit and relevant relationships among users based upon the patterns of their existing interactions. Our work is grounded in the context of item recommendations on Amazon. We investigate whether users' reply patterns can be used to identify these meaningful relationships and show that different degrees of relationships do exist. We develop global measures of relationship strength and observe that users tend to form strong connections when they are evaluating subjective items such as books and movies. We then design a probabilistic mechanism to distinguish meaningful connections from connections formed by chance and extract implicit communities. We finally show that these communities can be used for hybrid recommender systems that improve recommendations over existing collaborative filtering approaches.",
keywords = "implicit communities, recommendation, recommender system, social network",
author = "Euijin Choo and Ting Yu and Min Chi and Yan Sun",
year = "2014",
month = "1",
day = "1",
doi = "10.1145/2600057.2602906",
language = "English",
isbn = "9781450325653",
pages = "489--506",
booktitle = "EC 2014 - Proceedings of the 15th ACM Conference on Economics and Computation",
publisher = "Association for Computing Machinery",

}

TY - GEN

T1 - Revealing and incorporating implicit communities to improve recommender systems

AU - Choo, Euijin

AU - Yu, Ting

AU - Chi, Min

AU - Sun, Yan

PY - 2014/1/1

Y1 - 2014/1/1

N2 - Social connections often have a significant influence on personal decision making. Researchers have proposed novel recommender systems that take advantage of social relationship information to improve recommendations. These systems, while promising, are often hindered in practice. Existing social networks such as Facebook are not designed for recommendations and thus contain many irrelevant relationships. Many recommendation platforms such as Amazon often do not permit users to establish explicit social relationships. And direct integration of social and commercial systems raises privacy concerns. In this paper we address these issues by focusing on the extraction of implicit and relevant relationships among users based upon the patterns of their existing interactions. Our work is grounded in the context of item recommendations on Amazon. We investigate whether users' reply patterns can be used to identify these meaningful relationships and show that different degrees of relationships do exist. We develop global measures of relationship strength and observe that users tend to form strong connections when they are evaluating subjective items such as books and movies. We then design a probabilistic mechanism to distinguish meaningful connections from connections formed by chance and extract implicit communities. We finally show that these communities can be used for hybrid recommender systems that improve recommendations over existing collaborative filtering approaches.

AB - Social connections often have a significant influence on personal decision making. Researchers have proposed novel recommender systems that take advantage of social relationship information to improve recommendations. These systems, while promising, are often hindered in practice. Existing social networks such as Facebook are not designed for recommendations and thus contain many irrelevant relationships. Many recommendation platforms such as Amazon often do not permit users to establish explicit social relationships. And direct integration of social and commercial systems raises privacy concerns. In this paper we address these issues by focusing on the extraction of implicit and relevant relationships among users based upon the patterns of their existing interactions. Our work is grounded in the context of item recommendations on Amazon. We investigate whether users' reply patterns can be used to identify these meaningful relationships and show that different degrees of relationships do exist. We develop global measures of relationship strength and observe that users tend to form strong connections when they are evaluating subjective items such as books and movies. We then design a probabilistic mechanism to distinguish meaningful connections from connections formed by chance and extract implicit communities. We finally show that these communities can be used for hybrid recommender systems that improve recommendations over existing collaborative filtering approaches.

KW - implicit communities

KW - recommendation

KW - recommender system

KW - social network

UR - http://www.scopus.com/inward/record.url?scp=84903130197&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84903130197&partnerID=8YFLogxK

U2 - 10.1145/2600057.2602906

DO - 10.1145/2600057.2602906

M3 - Conference contribution

SN - 9781450325653

SP - 489

EP - 506

BT - EC 2014 - Proceedings of the 15th ACM Conference on Economics and Computation

PB - Association for Computing Machinery

ER -