In e-commerce websites a wide verity of products are available. Additionally, in large e-commerce websites every day a millions of products are introduced. Among them not all the products are very good in quality and services. On the other hand, the clients are mainly depends on product review and rating given on product review and rating section. Therefore the product purchasing or buying decisions are significantly depends on the product rating and review. It is a highly influential factor and can mislead the buyer’s decision. In this paper, we proposed a spam review identification model for helping the users by providing spam free and actual reviews of the product. The proposed model will use a text feature extraction technique using TF-IDF and chi-square test. Then, an Artificial Neural Network is trained for identifying the spam product reviews. In order, to conduct the experiments Amazon product review dataset has been used. Additionally a comparative performance study has also been carried out with SVM based classifier to justify the proposed model. According to the obtained performance we have found the proposed model is far superior than the SVM based spam detection model.

e-commerce, product review, cyber security, fraud, text classification, spam review.