Twitter Sentiment Analysis using Supervised Machine Learning Techniques

Along with the evolution of time the large number of people used the social media platform to share views. This makes more people can communicate with each other. Alongside these benefits, it has some negative sides also which brings animosity towards some part of individuals. It can also include the hate speech. Hate speech is the speech that might include the abusive or threatening words which affects the community. Such type of speech need to be detected and removed from social media platform before spreading. Analysis of sentiments is the method of deciding whether the sentiments in the text is hatred or not hatred. We analyzed the Twitter dataset using weka software. In this dataset there were total 5000 Tweets and we applied two filters(Tweet to Sparse Feature Vector, Tweet to Lexicon Feature Vector) on it to give model accuracy of machine learning. The experimental result in both cases of Twitter dataset has an highest accuracy of Random forest technique i.e.93%.