Abstract:

In recent days, the businesses are running based on E-commerce technologies. The customers are purchasing the products by considering the previous reviews given by other persons. The problem here is some times the customers are feeling that sellers are given fake reviews to gain the fame of their products and increasing the profits. To solve this problem of spam reviews detection several machine learning techniques like SVM, Extreme learning machine , decision tree and other methods are used. By using these methods some issues like imbalance data and unlabeled data are arrived. To solve these problems we use deep learning in spam detection. We proposed the Random forest(RF) and Recurrent neural networks(RNN) with attention mechanism for solving the above mentioned imbalanced data problems. We conducted experiments of proposed methods with other existing machine learning techniques on Amazon reviews data set with evaluated using measures like accuracy, AUC and F1score.Our proposed method given the better evaluation results than other mentioned machine learning methods

Description:

.