Fake Reviews Detection Using Machine Learning

Swijel Dmello, Ridhi Bauskar, Apoorva Shet

Abstract: Fake Review Detection is a very crucial task in the field of e-commerce and business. It plays a pivoting role in the decision-making and quality assessment of a product. However, this task of reviewing is carried out manually by humans or by a random review checker mechanism. In order to review false positive and false negative reviews, a fake review dataset has been used. Due to the advancements in the field of Machine Learning and Natural Language Processing, these algorithms could be leveraged to detect fake reviews with high accuracy and in a short amount of time which would save a huge amount of manual effort. This paper proposes an approach to detecting fake reviews through the various advanced machine learning techniques like Support Vector Machines, Decision Trees, etc. The system put forward is a web-based solution that provides an accurate result whether the given review is valid or not.

Keywords: Fake Review, Sentiment Analysis, Machine learning, SVM, KNN, GBB, NLTK, SGD, Decision Tree, Logistic Regression, MNB