Website Evaluation Using Opinion Mining

In this modern era everyone is aware of comment analysis. For example, if a person wants to buy a product, the first thing will do is to see the review of that particular product, based on review they choosing the product. By this way the reviews made a huge impact on the products and on websites. There are number of websites increasing in the internet, which provides various facilities like shopping, trading, online betting etc. Some of them are good and some are questionable like whether the website is beneficial or not. Many of them are facing the issues when choosing a website, in order to overcome some of these fraud websites, we proposed a system that evaluates the websites based on the reviews or feedback provided by the customer. For this we collected the twitter dataset because on twitter platform various number of emotions are expressed on the products, user interface, online transactions etc. as well as we used different kinds of classifiers for classifying the text such are Naive-Bayes, Random forest, Decision tree, Support Vector Machine, K Nearest Neighbor's, Logistic Regression. In this paper we utilize the Logistic Regression classifier is used for classification of the text and for generating the sentiment of the word, which achieves the better accuracy (99.50%), recall and F1 score among the other classifiers used in this paper and naïve-bayes has achieved the best high precision rather among the classifiers used. This paper emphasizes that the critiques expressed in the form of reviews are used to analyze the behavior is positive, negative or neutral and the system is used to rank satisfactory classifiers for sentiments.