Fake Reviews Detection using Supervised Machine Learning

Authors

  • Dr. K. Uday Kumar Reddy, P. Sukesh Kumar Reddy, K. Venkata Sunanda, T. Sree Lakshmi, M. Venkata Bhargav Kumar Reddy

Abstract

With the continuous development of e-commerce systems, online reviews are often seen as an important factor in building and maintaining a good reputation. Apart from that, they play an effective role for users of the decision-making process at the end of life. Usually a positive review for a target item attracts more customers and lead to a large increase in turnover. Of our time, misleading or false reviews are written intentionally to create reputation and attracting potential customers. So the identification fake reviews are a living and ongoing area of research. Identification fake reviews don't just depend on the main features of the reviews but also on the behavior of the examiners. This paper offers a machine learning approach to identify false assessments. Moreover on the process of extracting features from magazines, this article applies different technical functions to extract different behaviors examiners. The document compares the performance of different experiments conducted on a real Yelp dataset of restaurant reviews with and without features derived from user behavior. In in both cases we compare the performance of different classifiers; KNN, Naive Bayes (NB), SVM, logistic and random regression Forest. Also various n-gram language models in particular bigram and trigram are taken into account during the reviews. The results show that KNN (K=7) outperforms the rest of the ratings in terms of f-score that achieves the best f-score 82.40%. The results show that the f-score increased by 3.80% when considering behavioral traits extracted from reviewers consideration.

Published

2021-09-11

How to Cite

Dr. K. Uday Kumar Reddy, P. Sukesh Kumar Reddy, K. Venkata Sunanda, T. Sree Lakshmi, M. Venkata Bhargav Kumar Reddy. (2021). Fake Reviews Detection using Supervised Machine Learning. Drugs and Cell Therapies in Hematology, 10(3), 1132–1139. Retrieved from http://www.dcth.org/index.php/journal/article/view/1022

Issue

Section

Articles