Online Review Spam Detection by New Linguistic Features

Karami, Amir; Zhou, Bin

Permalink

https://hdl.handle.net/2142/73749 Copy

Description

Title

Online Review Spam Detection by New Linguistic Features

Author(s)

• Karami, Amir
• Zhou, Bin

Issue Date

2015-03-15

Keyword(s)

• data analytics and evaluation
• text/data/knowledge mining

Abstract

With the fast growing and importance of online reviews, malicious users start to abuse the online review websites and deliberately post low quality, untrustworthy, or even fraudulent reviews, which are typically referred to as ``spam reviews''. Many existing studies on review spam detection are based on classification models. Features such as the number of verbs used in the reviews are commonly used to construct the spam review classification model. Surprisingly, many linguistic features of users' reviews have not been thoroughly considered for review spam detection. In this paper, we focus on different types of linguistic features and evaluate their performance on detecting spam reviews. Our empirical evaluation conducted on a spam review benchmark dataset validated the proposed features significantly improve the performance of online review spam detection, reaching more than 93\% accuracy.

Publisher

iSchools

Series/Report Name or Number

iConference 2015 Proceedings

Type of Resource

text

Language

English

Permalink

http://hdl.handle.net/2142/73749

Copyright and License Information

Copyright 2015 is held by the authors. Copyright permissions, when appropriate, must be obtained directly from the authors.

Owning Collections