Prediction and Analysis of Sentiments on Twitter Data using Hybrid Naive Bayes Approach
Ch. Srinivasa Rao1, G. Satyanarayana Prasad2, Vedula Venkateswara Rao3

1Ch Srinivasa Rao, Research Scholar, Dept of CSE Acharya Nagarjuna University, Guntur, Associate Professor, Dept of CS, SVKP & Dr K S RAJU A&Sc College, Penugonda, (A.P), India.
2Dr. G Satyanarayana Prasad, Professor, Dept of CSE, Dean, Training & Placements, RVR & JC College of Engineering Chowdavaram, Guntur, (A.P), India.
3Dr. Vedula Venkateswara Rao, Professor, Dept of CSE,Sri Vasavi Engineering College, Pedatadepalli, Tadepalligudem, (A.P). India, Bags of the wholesale user created content that have been premeditated are blogs [1] and product/movie [2] reviews.

Manuscript received on 02 June 2019 | Revised Manuscript received on 10 June 2019 | Manuscript published on 30 June 2019 | PP: 1087-1091 | Volume-8 Issue-8, June 2019 | Retrieval Number: H6583068819/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (

Abstract: Every single day, billions of users places their opinions on innumerable aspects of our life and politics with use of micro blogging above the internet. Microbloging websites are prosperous cradles of data for confidence mining and sentiment analysis. In our research work, we emphasis on expending Twitter for sentiment analysis for mining opinions about events, products, individuals and consume it for accepting the current tendencies. Twitter permits its customers a restriction of merely 140 letterings; this limitation powers the customer to be crisp as well as sensitive at the same time. This eventually creates twitter awater of sentiments. Twitter also offers designer responsive streaming. We scurry datasets above5 million tweets by a conventional intended crawler for sentiment analysis tenacity. We suggest a hybrid naive bayes classifier by means of assimilating an English lexical dictionary called SentiWordNet to the prevailing machine learning naïve bayes classifier algorithm. Hybrid naive bayes categorizes the tweets in negative and positive groups independently. Investigational results validated the dominance of hybrid naive bayes on multi sized datasets comprising of assortment of keywords over prevailing methods producing more than 90%correctness in common and 98.29%correctness in the best case. In our research work, we executed through English; nevertheless, the recommended method can be applied with any new language, as long as that language lexicon dictionary.
Keyword: Micro-blog, Predictions, Sentiment mining, Twitter.
Scope of the Article: Pattern Recognition and Analysis.