1 / 29

Reputation Systems

Reputation Systems. Guest Lecture Paul Resnick Associate Professor Univ. of Michigan School of Information presnick@umich.edu. Learning Objectives. Understand What a reputation system is Theory about when and why it should work Open research questions Participate in design

schuyler
Download Presentation

Reputation Systems

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Reputation Systems Guest Lecture Paul Resnick Associate Professor Univ. of Michigan School of Information presnick@umich.edu

  2. Learning Objectives • Understand • What a reputation system is • Theory about when and why it should work • Open research questions • Participate in design • Recognize situations when it might be helpful • Raise some of the difficult design challenges

  3. Outline • What is a reputation system? • Theory: when/why they should work • Empirical results • Design space • Case study: NPAssist recommender

  4. Definition • A Reputation System… • Collects • Distributes • Aggregates • …information about behavior

  5. Examples • BBB • Bizrate • eBay • Expertise sites • Epinions “top reviewers” • Slashdot karma system

  6. Why Reputation Systems • Interacting with strangers • Sellers (Exchange Partners) Vary • Skill • Effort • Ethics

  7. Other Trust-Inducing Mechanisms in E-commerce • Insurance • Escrow • Fraud Prosecution

  8. How Reputation Systems Should Work • Information • Incentive • Self-selection

  9. Some Issues • Anonymity • Name changes • Name trades • Lending reputations • Eliciting evaluation • Honesty of evaluations

  10. Anonymity Analysis

  11. 1L Pseudonyms • Third-party issues pseudonyms • No cost • Not replaceable • Reveal name to third party • Don’t reveal mapping of name to pseudonym

  12. Empirical Results: eBay • Feedback is provided • It’s almost all positive • Reputations are informative • Reputation benefits • Effect on probability of sale • Effect on price

  13. Provision of Feedback • Negatives: paid but did not receive; seller cancelled; not as advertised; communication • Neutrals: slow shipping, not as advertised, communication

  14. Feedback Profiles of Buyers and Sellers

  15. Predicting Problematic Transactions • Logistic Regressionf(0,0) = 1.91% f(100,0) = .18% f(100,3) = .53% N = 36233 Beginning Block Number 0. Initial Log Likelihood Function -2 Log Likelihood 2194.3468 -2 Log Likelihood 2075.420 Dependent Variable.. NEGNEUT ---------------------- Variables in the Equation ----------------------- Variable B S.E. Wald df Sig R Exp(B) LNNPOS .7712 .1179 42.7907 1 .0000 .1363 2.1624 LNPOS -.5137 .0475 116.8293 1 .0000 -.2288 .5983 Constant -3.9399 .1291 931.3828 1 .0000

  16. Predictive Value

  17. Some Recently Completed Work • Experiment: does reputation affect profit? • Many positives: Yes, but only a little (8.1%) • One or two negatives: No • Incentives for quality feedback provision • Can pay based on agreement among raters

  18. Studies Currently Underway • Feedback provision (empirical) • Reciprocation, altruism, and free riding • Dynamics: learning and selection (empirical) • Geography: trust and trustworthiness by state

  19. Design Space • Rating scales • Aggregation of ratings • Who rates? • Incentives for raters • Identification/Anonymity • Exchange partners • Evaluation providers

  20. Case Study • Goal: help Michigan non-profits select consultants and other service providers • Is this a good candidate for a reputation system?

  21. Case Study • Goal: help Michigan non-profits select consultants and other service providers • Is this a good candidate for a reputation system? • Interacting with strangers • Sellers (Exchange Partners) Vary • Skill • Effort • Ethics

  22. Case Study Design Choices • Rating scales • Aggregation of ratings • Who rates? • Incentives for raters • Identification/Anonymity • Exchange partners • Evaluation providers

  23. Summary • RS inform, incent, select • Opportunity for RS: interactions with strangers • Design space • Scales, aggregation, raters, incentives, anonymity

More Related