Skip to main navigation Skip to search Skip to main content

Attacking naive bayes journal recommendation systems

  • Sheng Wang
  • , Mohan Li
  • , Yinyin Cai
  • , Zhaoquan Gu*
  • *Corresponding author for this work
  • Guangzhou University

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Recommendation systems have been extensively adopted in various applications. However, with the security concern of artificial intelligence, the robustness of such systems against malicious attacks has been studied in recent years. In this paper, we build a journal recommendation system based on the Naive Bayesian algorithm which helps recommend suitable journals for the authors. Since journal recommendation systems may also suffer from various attacks, we explore attack methods on the malicious data. We construct specific malicious data to attack the availability of training data, and such deviations in the training data could lead to poor recommendation accuracy. We also conduct extensive experiments and the results show that the recommendation accuracy could be dramatically reduced under such attacks.

Original languageEnglish
Title of host publicationArtificial Intelligence and Security - 6th International Conference, ICAIS 2020, Proceedings
EditorsXingming Sun, Jinwei Wang, Elisa Bertino
PublisherSpringer Science and Business Media Deutschland GmbH
Pages118-128
Number of pages11
ISBN (Print)9789811581007
DOIs
StatePublished - 2020
Externally publishedYes
Event6th International Conference on Artificial Intelligence and Security,ICAIS 2020 - Hohhot, China
Duration: 17 Jul 202020 Jul 2020

Publication series

NameCommunications in Computer and Information Science
Volume1254 CCIS
ISSN (Print)1865-0929
ISSN (Electronic)1865-0937

Conference

Conference6th International Conference on Artificial Intelligence and Security,ICAIS 2020
Country/TerritoryChina
CityHohhot
Period17/07/2020/07/20

Keywords

  • Journal recommendation systems
  • Malicious data attack
  • Naive Bayes

Fingerprint

Dive into the research topics of 'Attacking naive bayes journal recommendation systems'. Together they form a unique fingerprint.

Cite this