Sentiment analysis on twitter social media product reviews

Abstract

Well timed and profitable analytics over “Social Networks” is now a lead for gain in many industries to increase brand exposure and broaden customer reach by analyzing the reviews and recommending products to consumers. The objective is to extract the unstructured conversation or posts from Twitter and present a methodology for sentiment analysis on Twitter that reviews the products by considering the consumer’s sentiments Naive Bayesian Classifier. Samsung smartphones were taken into analysis and the experiment reveals six possible different sentiments: joy, surprise, anger, disgust, fear and sadness in the analysing process of the tweets. From the experimental results, it is observed that 80% of the customers has given positive feedback on the product and 20% has given negative feedback and the level of accuracy of the proposed approach is 90%.