Finding Valuable Yelp Comments by Personality, Content, Geo, and Anomaly Analysis

Author

Koven, Jay ; Siadati, Hossein ; Ching-Yung Lin

Author_Institution

Polytech. Sch. of Eng., NYU, New York, NY, USA

fYear

2014

fDate

14-14 Dec. 2014

Firstpage

1215

Lastpage

1218

Abstract

User reported experiences and opinions are used by peers to make decisions about where to go and what to buy. Unfortunately, not all users or opinions are honest. Many opinions are fabricated and may be submitted by automated systems or by people who are recruited by businesses and search engine optimizers to write good reviews. Such reviews and ratings are called spam reviews. These are misleading for users and troublesome for honest businesses. While most current efforts to tackle this problem are focused on spam review detection, in this paper we focused on detecting authentic and valuable reviews for a front-end application that reorders the reviews. In this manner, we have identified several features based upon the content of the reviews as well as identifying behavioural features of reviewers to pinpoint useful reviews with 80% accuracy.

Keywords

Web sites; information analysis; pattern classification; unsolicited e-mail; Yelp comments; anomaly analysis; content analysis; geo analysis; personality analysis; spam review detection; user experience; user opinion; Accuracy; Business; Data mining; Educational institutions; Feature extraction; Machine learning algorithms; Vectors; Machine Learning; Security; Usability; Valuable Review Detection;

fLanguage

English

Publisher

ieee

Conference_Titel

Data Mining Workshop (ICDMW), 2014 IEEE International Conference on

Conference_Location

Shenzhen

Print_ISBN

978-1-4799-4275-6

Type

conf

DOI

10.1109/ICDMW.2014.49

Filename

7022737