Abstract:
It is important for future customers to make choices on the basis of online feedback.
The utility, though, gives rise to a curse - a false opinion spam. Deceptive opinion
spam misleads prospective consumers and organisations to reshape their companies
and inhibits opinion-mining strategies from drawing correct conclusions. Thus, the
identification of misleading feedback has become more and more forceful. In this
project, we try to figure out how to differentiate between fake reviews and genuine
by using the linguistic features of the Yelp Filter Dataset. We have
suggested an approach for features extraction dependent on the Latent Dirichlet
Allocation (LDA). The findings of the experiment have shown that the procedure is
efficient. The growing prevalence of online reviews also encourages the false review
writing industry, which relates to paying human writers creating disappointing
reviews to manipulate the opinions of readers. Our project solves this issue by
developing a classifier that takes the evaluation text and its reviewer s specific data as
inputs and outputs ifthe review is valid.
