Detection of opinion spam based on anomalous rating deviation

作者：

Highlights：

• We propose a novel approach for detecting opinion spam based on reviewer ratings, with no need for text analysis.

• Our approach identifies reviewers having a significant number of reviews that disagree with majority opinion.

• Our approach requires significantly less running time than other, comparable approaches.

• Our approach is shown to successfully identify opinion spammers in synthetic and real world data sets.

摘要

•We propose a novel approach for detecting opinion spam based on reviewer ratings, with no need for text analysis.•Our approach identifies reviewers having a significant number of reviews that disagree with majority opinion.•Our approach requires significantly less running time than other, comparable approaches.•Our approach is shown to successfully identify opinion spammers in synthetic and real world data sets.

论文关键词：Anomaly detection,Binomial regression,Classification,Online product reviews,Opinion spam,Review spam

论文评审过程：Received 7 April 2015, Revised 19 June 2015, Accepted 11 July 2015, Available online 21 July 2015, Version of Record 29 August 2015.

论文官网地址：https://doi.org/10.1016/j.eswa.2015.07.019