Understanding longitudinal dynamics of recommender systems performance

An agent-based modeling approach

Gediminas Adomavicius, Alok Gupta, Wolfgang Ketter, Jingjing Zhang

Research output: Contribution to conferencePaper

Abstract

We develop a general agent-based modeling and simulation approach to study the impact of various factors on the temporal dynamics of recommender systems' performance. In this paper, we specifically focus on exploring the product consumption strategies and show that, in the long run, they can significantly impact the system's recommendation quality in various ways. For example, the more heavily users rely on the system's recommendations to make item choices, the more inaccurate and less diverse (and, thus, arguably less useful) system's predictions become in the future, which has important implications for recommender systems design. More generally, the proposed agent-based simulation approach allows for comprehensive analysis of longitudinal recommender systems performance under a variety of diverse conditions, which typically is not feasible with live real-world systems.

Original languageEnglish (US)
StatePublished - Jan 1 2013
Event23rd Workshop on Information Technology and Systems: Leveraging Big Data Analytics for Societal Benefits, WITS 2013 - Milan, Italy
Duration: Dec 14 2013Dec 15 2013

Other

Other23rd Workshop on Information Technology and Systems: Leveraging Big Data Analytics for Societal Benefits, WITS 2013
CountryItaly
CityMilan
Period12/14/1312/15/13

Fingerprint

Recommender systems
Systems analysis

Keywords

  • Agent-based simulation
  • Dynamics of recommender systems
  • Item consumption patterns

Cite this

Adomavicius, G., Gupta, A., Ketter, W., & Zhang, J. (2013). Understanding longitudinal dynamics of recommender systems performance: An agent-based modeling approach. Paper presented at 23rd Workshop on Information Technology and Systems: Leveraging Big Data Analytics for Societal Benefits, WITS 2013, Milan, Italy.

Understanding longitudinal dynamics of recommender systems performance : An agent-based modeling approach. / Adomavicius, Gediminas; Gupta, Alok; Ketter, Wolfgang; Zhang, Jingjing.

2013. Paper presented at 23rd Workshop on Information Technology and Systems: Leveraging Big Data Analytics for Societal Benefits, WITS 2013, Milan, Italy.

Research output: Contribution to conferencePaper

Adomavicius, G, Gupta, A, Ketter, W & Zhang, J 2013, 'Understanding longitudinal dynamics of recommender systems performance: An agent-based modeling approach' Paper presented at 23rd Workshop on Information Technology and Systems: Leveraging Big Data Analytics for Societal Benefits, WITS 2013, Milan, Italy, 12/14/13 - 12/15/13, .
Adomavicius G, Gupta A, Ketter W, Zhang J. Understanding longitudinal dynamics of recommender systems performance: An agent-based modeling approach. 2013. Paper presented at 23rd Workshop on Information Technology and Systems: Leveraging Big Data Analytics for Societal Benefits, WITS 2013, Milan, Italy.
Adomavicius, Gediminas ; Gupta, Alok ; Ketter, Wolfgang ; Zhang, Jingjing. / Understanding longitudinal dynamics of recommender systems performance : An agent-based modeling approach. Paper presented at 23rd Workshop on Information Technology and Systems: Leveraging Big Data Analytics for Societal Benefits, WITS 2013, Milan, Italy.
@conference{8cba169974ca4f0e9f4c366c0c78764b,
title = "Understanding longitudinal dynamics of recommender systems performance: An agent-based modeling approach",
abstract = "We develop a general agent-based modeling and simulation approach to study the impact of various factors on the temporal dynamics of recommender systems' performance. In this paper, we specifically focus on exploring the product consumption strategies and show that, in the long run, they can significantly impact the system's recommendation quality in various ways. For example, the more heavily users rely on the system's recommendations to make item choices, the more inaccurate and less diverse (and, thus, arguably less useful) system's predictions become in the future, which has important implications for recommender systems design. More generally, the proposed agent-based simulation approach allows for comprehensive analysis of longitudinal recommender systems performance under a variety of diverse conditions, which typically is not feasible with live real-world systems.",
keywords = "Agent-based simulation, Dynamics of recommender systems, Item consumption patterns",
author = "Gediminas Adomavicius and Alok Gupta and Wolfgang Ketter and Jingjing Zhang",
year = "2013",
month = "1",
day = "1",
language = "English (US)",
note = "23rd Workshop on Information Technology and Systems: Leveraging Big Data Analytics for Societal Benefits, WITS 2013 ; Conference date: 14-12-2013 Through 15-12-2013",

}

TY - CONF

T1 - Understanding longitudinal dynamics of recommender systems performance

T2 - An agent-based modeling approach

AU - Adomavicius, Gediminas

AU - Gupta, Alok

AU - Ketter, Wolfgang

AU - Zhang, Jingjing

PY - 2013/1/1

Y1 - 2013/1/1

N2 - We develop a general agent-based modeling and simulation approach to study the impact of various factors on the temporal dynamics of recommender systems' performance. In this paper, we specifically focus on exploring the product consumption strategies and show that, in the long run, they can significantly impact the system's recommendation quality in various ways. For example, the more heavily users rely on the system's recommendations to make item choices, the more inaccurate and less diverse (and, thus, arguably less useful) system's predictions become in the future, which has important implications for recommender systems design. More generally, the proposed agent-based simulation approach allows for comprehensive analysis of longitudinal recommender systems performance under a variety of diverse conditions, which typically is not feasible with live real-world systems.

AB - We develop a general agent-based modeling and simulation approach to study the impact of various factors on the temporal dynamics of recommender systems' performance. In this paper, we specifically focus on exploring the product consumption strategies and show that, in the long run, they can significantly impact the system's recommendation quality in various ways. For example, the more heavily users rely on the system's recommendations to make item choices, the more inaccurate and less diverse (and, thus, arguably less useful) system's predictions become in the future, which has important implications for recommender systems design. More generally, the proposed agent-based simulation approach allows for comprehensive analysis of longitudinal recommender systems performance under a variety of diverse conditions, which typically is not feasible with live real-world systems.

KW - Agent-based simulation

KW - Dynamics of recommender systems

KW - Item consumption patterns

UR - http://www.scopus.com/inward/record.url?scp=84907414776&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84907414776&partnerID=8YFLogxK

M3 - Paper

ER -