Proficient Machine Learning based Scheme for Classifying User Reviews

Abstract

Abstract: Each user represents his view on a product or issue using several aspects or features. A user’s opinion may be conveyed based on diverse levels of positivity or negativity. In this paper, a proficient technique is propounded for categorizing customer reviews taken from Twitter dataset. The features are extracted using Bag of Words (BoW) algorithm and the information is categorized using modified Support Vector Machine (SVM) and ensemble classification. Further, the results are optimized using BAT algorithm which improves the classification accuracy. It is notable that the scheme proposed offers better accuracy, precision, recall and involves less time.