De-biasing user preference ratings in recommender systems completed research paper

Gediminas Adomavicius, Jesse C Bockstedt, Shawn Curley, Jingjing Zhang

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

Abstract

Prior research has shown that online recommendations can lead to significant distortion of users' preference ratings and economic behavior. Specifically, the self-reported preference rating that is submitted by a user to a recommender system can be affected by the previously observed system's recommendation. This research explores two approaches to removing anchoring biases from self-reported consumer ratings. The first proposed approach is based on a computational post-hoc de-biasing algorithm that systematically adjusts the user-submitted ratings that are known to be biased. The second approach is a user-interface-driven solution that tries to minimize anchoring biases at rating collection time. Our empirical investigation explicitly demonstrates the impact of biased vs. unbiased ratings on recommender systems' predictive performance. It also indicates that the post-hoc algorithmic de-biasing approach is very problematic, most likely due to the fact that the anchoring effects can manifest themselves very differently for different users and items. This further emphasizes the importance of proactively avoiding anchoring biases at the time of rating collection. Further, through laboratory experiments, we demonstrate that certain interface designs of recommender systems are more advantageous than others in effectively reducing anchoring biases.

Original languageEnglish (US)
Title of host publication24th Workshop on Information Technology and Systems
PublisherUniversity of Auckland Business School
StatePublished - 2014
Event24th Annual Workshop on Information Technologies and Systems: Value Creation from Innovative Technologies, WITS 2014 - Auckland, New Zealand
Duration: Dec 17 2014Dec 19 2014

Other

Other24th Annual Workshop on Information Technologies and Systems: Value Creation from Innovative Technologies, WITS 2014
CountryNew Zealand
CityAuckland
Period12/17/1412/19/14

Fingerprint

Recommender systems
User interfaces
Economics
Experiments

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Information Systems
  • Computer Science Applications

Cite this

Adomavicius, G., Bockstedt, J. C., Curley, S., & Zhang, J. (2014). De-biasing user preference ratings in recommender systems completed research paper. In 24th Workshop on Information Technology and Systems University of Auckland Business School.

De-biasing user preference ratings in recommender systems completed research paper. / Adomavicius, Gediminas; Bockstedt, Jesse C; Curley, Shawn; Zhang, Jingjing.

24th Workshop on Information Technology and Systems. University of Auckland Business School, 2014.

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

Adomavicius, G, Bockstedt, JC, Curley, S & Zhang, J 2014, De-biasing user preference ratings in recommender systems completed research paper. in 24th Workshop on Information Technology and Systems. University of Auckland Business School, 24th Annual Workshop on Information Technologies and Systems: Value Creation from Innovative Technologies, WITS 2014, Auckland, New Zealand, 12/17/14.
Adomavicius G, Bockstedt JC, Curley S, Zhang J. De-biasing user preference ratings in recommender systems completed research paper. In 24th Workshop on Information Technology and Systems. University of Auckland Business School. 2014
Adomavicius, Gediminas ; Bockstedt, Jesse C ; Curley, Shawn ; Zhang, Jingjing. / De-biasing user preference ratings in recommender systems completed research paper. 24th Workshop on Information Technology and Systems. University of Auckland Business School, 2014.
@inproceedings{fcfebe4ab6304f9dad119cdf942f5718,
title = "De-biasing user preference ratings in recommender systems completed research paper",
abstract = "Prior research has shown that online recommendations can lead to significant distortion of users' preference ratings and economic behavior. Specifically, the self-reported preference rating that is submitted by a user to a recommender system can be affected by the previously observed system's recommendation. This research explores two approaches to removing anchoring biases from self-reported consumer ratings. The first proposed approach is based on a computational post-hoc de-biasing algorithm that systematically adjusts the user-submitted ratings that are known to be biased. The second approach is a user-interface-driven solution that tries to minimize anchoring biases at rating collection time. Our empirical investigation explicitly demonstrates the impact of biased vs. unbiased ratings on recommender systems' predictive performance. It also indicates that the post-hoc algorithmic de-biasing approach is very problematic, most likely due to the fact that the anchoring effects can manifest themselves very differently for different users and items. This further emphasizes the importance of proactively avoiding anchoring biases at the time of rating collection. Further, through laboratory experiments, we demonstrate that certain interface designs of recommender systems are more advantageous than others in effectively reducing anchoring biases.",
author = "Gediminas Adomavicius and Bockstedt, {Jesse C} and Shawn Curley and Jingjing Zhang",
year = "2014",
language = "English (US)",
booktitle = "24th Workshop on Information Technology and Systems",
publisher = "University of Auckland Business School",

}

TY - GEN

T1 - De-biasing user preference ratings in recommender systems completed research paper

AU - Adomavicius, Gediminas

AU - Bockstedt, Jesse C

AU - Curley, Shawn

AU - Zhang, Jingjing

PY - 2014

Y1 - 2014

N2 - Prior research has shown that online recommendations can lead to significant distortion of users' preference ratings and economic behavior. Specifically, the self-reported preference rating that is submitted by a user to a recommender system can be affected by the previously observed system's recommendation. This research explores two approaches to removing anchoring biases from self-reported consumer ratings. The first proposed approach is based on a computational post-hoc de-biasing algorithm that systematically adjusts the user-submitted ratings that are known to be biased. The second approach is a user-interface-driven solution that tries to minimize anchoring biases at rating collection time. Our empirical investigation explicitly demonstrates the impact of biased vs. unbiased ratings on recommender systems' predictive performance. It also indicates that the post-hoc algorithmic de-biasing approach is very problematic, most likely due to the fact that the anchoring effects can manifest themselves very differently for different users and items. This further emphasizes the importance of proactively avoiding anchoring biases at the time of rating collection. Further, through laboratory experiments, we demonstrate that certain interface designs of recommender systems are more advantageous than others in effectively reducing anchoring biases.

AB - Prior research has shown that online recommendations can lead to significant distortion of users' preference ratings and economic behavior. Specifically, the self-reported preference rating that is submitted by a user to a recommender system can be affected by the previously observed system's recommendation. This research explores two approaches to removing anchoring biases from self-reported consumer ratings. The first proposed approach is based on a computational post-hoc de-biasing algorithm that systematically adjusts the user-submitted ratings that are known to be biased. The second approach is a user-interface-driven solution that tries to minimize anchoring biases at rating collection time. Our empirical investigation explicitly demonstrates the impact of biased vs. unbiased ratings on recommender systems' predictive performance. It also indicates that the post-hoc algorithmic de-biasing approach is very problematic, most likely due to the fact that the anchoring effects can manifest themselves very differently for different users and items. This further emphasizes the importance of proactively avoiding anchoring biases at the time of rating collection. Further, through laboratory experiments, we demonstrate that certain interface designs of recommender systems are more advantageous than others in effectively reducing anchoring biases.

UR - http://www.scopus.com/inward/record.url?scp=84924674407&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84924674407&partnerID=8YFLogxK

M3 - Conference contribution

BT - 24th Workshop on Information Technology and Systems

PB - University of Auckland Business School

ER -