Sentiment analysis of user's views using machine learning

Sentiment Analysis or Opinion Mining is an important concept in today's world and due to the increased use of media it has become a huge source of database. Since everybody in the modern era is involved with some social media platform, the public mood is hugely reflected in the social media platform today. This study proposes to utilize this source of information and predict the all sentiments of public towards the food price in India expressed over twitter and twitter API is used for extracting live tweets. Oauth is used as handler and tweets are filtered for specific keywords and location using latitude and longitude data. The tweets are saved into a database. They first preprocessed for elimination of stop word, special characters, short words etc, after that stemming and tokenization steps are applied and TF-IDF score is calculated for all the keywords. A term document matrix (TDM) is created which is fed into the classifiers for classification. KNN and Naïve Baye's has been analyzed in this study and Hybrid algorithm using them was designed. The results of KNN and Naïve Baye's classifier in sentiment classification were found to be significant while the hybrid-KNN outperforms the Naïve Baye's Classifiers in terms of accuracy..