Deceptive Opinion Detection Using Machine Learning Techniques

Abstract

Nowadays, online reviews have become a valuable resource for customer decision making before purchasing a product. Research shows that most of the people look at online reviews before purchasing any product. So, customers reviews are now become a crucial part of doing business online. Since review can either promote or demote a product or a service, so buying and selling fake reviews turns into a profitable business for some people now a days. In the past few years, deceptive review detection has attracted significant attention from both the industrial organizations and academic communities. However, the issue remains to be a challenging problem due to the lack of labeled dataset for supervised learning and evaluation. Also, study shows that both the state of the art computational approaches and human readers acquire an error rate of about 35% to 48% in identifying fake reviews. This study thoroughly investigated and analyzed customers' online reviews for deception detection using different supervised machine learning methods and proposes a machine learning model using stochastic gradient descent algorithm for the detection of spam review. To reduce bias and variance, bagging and boosting approach was integrated into the model. Furthermore, to select the most appropriate features in the feature selection step, some rules using regular expression were also generated. Experiments on hotel review dataset demonstrate the effectiveness of the proposed approach.

Credibility Detection on Twitter News Using Machine Learning Approach

“…This work Ref. [30] examined and analyzed customer's online opinions for credibility detection using different machine learning techniques to choose the suitable functions in functions choice set specific rules. Using normal terms was provided.…”

Fuzzy artificial bee colony‐based CNN‐LSTM and semantic feature for fake product review classification

“…The Stylometric feature comprises of the total number of characters in each sentence, average words in each sentence, the total number of emotional words as well as punctuation marks utilized in the review. 6…”

“…Here, both positive and negative opinions are considered as valuable one and this permits the people having nasty thoughts that make it easy for the people to post their opinions and support their product. 6 That kind of people are referred to as opinion spammer and such activity is said to be known as opinion spamming. Numerous researches have been carried out in sentiment analysis but the principal concern is only on the algorithms and not on detecting the fake reviews.…”