Identifying and minimizing the effects of malicious behavior in SERF Public Deposited

http://ir.library.oregonstate.edu/concern/technical_reports/mk61rj19c

Descriptions

Attribute NameValues
Creator
Abstract or Summary
  • Collaborative filtering (CF) algorithms are used in a wide range of internet applications. However the chief objective of using CF algorithms across most of these applications is to discover items that might be of interest to its users. CF algorithms work by obtaining feedback from users on the items that they browse and utilize that feedback to suggest recommendations to other users with similar tastes. CF algorithms rely heavily on input provided by humans and thus it is vital to verify that this information is appropriate. In this paper, we analyze various mechanisms by which users can enter malicious data to a CF system called SERF (System for Electronic Recommendation Filtering). We explore how bad data can be propagated through the system and can be used to manipulate the quality of recommendations. We also explore some techniques to counter the effects of bad data on the system. We report the results of our experiment with two simulated systems - a reputation system that utilizes a user's agreement and disagreement history to predict the trust that can be attributed to a user and a word weighting scheme based on word co-occurrence.
Resource Type
Date Available
Date Issued
Series
Keyword
Subject
Rights Statement
Publisher
Peer Reviewed
Language
Replaces
Additional Information
  • description.provenance : Made available in DSpace on 2012-12-04T17:42:04Z (GMT). No. of bitstreams: 1 2005-126.pdf: 564370 bytes, checksum: 430325d15307ec530622e119798c7ca2 (MD5) Previous issue date: 2005
  • description.provenance : Approved for entry into archive by Laura Wilson(laura.wilson@oregonstate.edu) on 2012-12-04T17:42:04Z (GMT) No. of bitstreams: 1 2005-126.pdf: 564370 bytes, checksum: 430325d15307ec530622e119798c7ca2 (MD5)
  • description.provenance : Submitted by Laura Wilson (laura.wilson@oregonstate.edu) on 2012-12-04T17:41:11Z No. of bitstreams: 1 2005-126.pdf: 564370 bytes, checksum: 430325d15307ec530622e119798c7ca2 (MD5)

Relationships

In Administrative Set:
Last modified: 07/18/2017

Downloadable Content

Download PDF
Citations:

EndNote | Zotero | Mendeley

Items