Factual Product Recommendation System by Eliminating Fake Reviews using Machine Learning Techniques

Abstract

The E-Commerce platform is rapidly growing at an unprecedented rate around the world which attracts
people from all age groups to satisfy their shopping needs. It gives more satisfaction to their users as
compared to retail shopping due to few reasons like huge number of e-commerce stores, easy to find a
product on e-store, provides detailed information of the product and e-stores provides customers reviews
about the product. Recent research states that the user feedback plays a vital role in deciding the sales of
the product. The customers reviews are huge in volume, which consists of actual feedback from customers
and fake feedback from spammers. So, the extraction of genuine review information is a difficult task for
the customers, manufactures and others who interested in customer feedback. In this paper Factual Product
Recommendation System (FPRS) is introduced. FPRS is an effective technique to solve fake reviews
problem and provide genuine reviews. FPRS incorporates the Naive Bayes Classifier (NBC) and Latent
Dirichlet Allocation (LDA) to sanitize and extract the genuine reviews and also includes the Matrix
Factorization (MF) to provide products recommendations.