Sorry, you need to enable JavaScript to visit this website.

ReviewSec: A Tool for Online Review Analysis

Citation Author(s):
Submitted by:
Yongbo Zeng
Last updated:
23 February 2016 - 1:43pm
Document Type:
Research Manuscript


Online review plays an important role when people
are making decisions to purchase a product/service. It is shown
that the sellers can benefit from boosting the reviews of their
products/services, or downgrading the reviews of their competitors.
Dishonest behavior on reviews can seriously affect both buyers
and sellers. In this work, we propose an algorithm that contains
a two-step analysis to detect whether a product’s reviews have
been manipulated. The first step is called consistency analysis,
which detects the variation in rating values. In the second step,
we introduce a novel angle to detect dishonest reviews, called
Equal Rating Opportunity (ERO) principle. We propose the ERO
analysis using the ANOVA method. Furthermore, we develop a
web-based system, referred to as ReviewSec, to conduct realtime
on-demand review manipulation detection. The ReviewSec
system includes three modules: 1) crawler, which download
reviews data from e-commerce website; 2) detector, consisting
of the consistency analysis and ERO analysis; and 3) web-based
interface. We believe that with the assistance of the ReviewSec
system, online shoppers can understand product reviews in a
better way and thereby reduce the risk of being misled by
untruthful reviews.

3 users have voted: Yongbo Zeng, Yihai Zhu, Zoltan Safar