A comparative analysis of machine learning algorithms for fake news detection

Abstract

Nowadays, online buying has become a worldwide phenomenon. People frequently purchase items online, and many of e-commerce sites provide a review option for client feedback. People generally make purchasing decisions based on customer reviews that are already available. Some website owners may employ spammers to write false reviews in order to boost product sales. Many approaches have been proposed by researchers in the past to detect fraudulent reviews. However, there is a critical need to identify and analyze the best machine learning algorithm to detect fraudulent reviews. Therefore, in this study machine learning algorithms including support vector machine (SVM), Random Forest (RF), Logistic Regression (LR), Multi-layer perceptron (NN), Long Short-Term Memory (LSTM) and Decision Tree (DT) are compared. The comparison is done by comparing the results of evaluation parameters i.e. Accuracy, Precision, Recall and F1-Measure. Results of this study shows that, RF is the best algorithms for detecting fake reviews.