Comparative Analysis of Classifier Algorithms Based on Sentimental Reviews

Major part of business is moving toward online. E-commerce companies are growing increasingly and growing fast. Capacity and ranges of products are also increasing in almost all domain areas. With this, customers are also showing their keen interest in online marketing. Seeing the interest of customers, e-commerce companies are using several strategies to attract the customers. Machine learning techniques are playing a very vital role in increasing the business where sentiment analysis is frequently used. One key area which is of major importance to increase the business is the sentiments hidden in the product reviews made by customers. By understanding the sentiments, companies can work on user satisfaction which is of utmost value to the business. There is a need to find and explore the algorithms working in the background and find the best one. In this paper, different classifiers have been analyzed and tested for their accuracy and other parameters on weka platform. It has been found that logistic and random forest classifier among Bayes Network, Naive Bayes, NaiveBayesMultinomialText, SimpleLogistic, SMO, Lazy IBk, Lazy LBL, Bagging, and MultiClassClassifier is giving the highest accuracy of classification of sentimental reviews.