In e-commerce websites, users are utilizing product reviews to make decisions about a product purchase. But the fake and spam reviews of products are influencing buyers’ decisions. In this context, recently different approaches are developed for classifying spam reviews, but most of the authors are only considering only the review text which provides partial information to classify the reviews as spam or legitimate. In this paper, we are proposing a new weighted classification technique to deal with this problem. The proposed technique utilizes the reviewer’s attributes and other people's opinions about the review for providing a score for the review. On the other hand, the review text is classified using the text classification approach to decide the initial review class. Further, a weight is calculated using the calculated score and initial class label. Using this weight we have decided on the final class label about the product class i.e. spam or legitimate. The experiments are carried out using the Amazon product review dataset and the performance of the model has been measured. The performance of the prepared model is measured and compared in terms of Precision, Recall, and F1-score. The comparison shows the proposed method provides more accurate results as compared to only text classification-based techniques.

Text mining, sentiment analysis, Spam Classification, weighted classifier, ecommerce, decision making process.