Effective Approaches for Classification and Rating of Users Reviews

Abstract

Organizations provide a platform for the users to express their opinion on the product in the form of reviews. Spam reviews are irrelevant reviews that mislead the consumers. In this paper we discuss semantic and machine learning approaches to classify reviews of trending products and provide rating to the reviews. We have used semantic and machine learning algorithms on five different products’ dataset. We have collected datasets comprising of both spam and non-spam reviews for training and testing purposes. We have obtained an average accuracy of 82.2% for the classification and 84.4% for review rating considering all the five products using Semantic approach. Similarly, we have obtained an accuracy of 82.2% using machine learning for the classification. For rating the review, we have obtained accuracy of 89.4% using machine learning. We found that both semantic and machine learning approaches perform well for classification of reviews. However for rating of reviews we found machine learning approach performed marginally better than semantic approach.