An Enhanced Technique for Analyzing Sentiments of Public Reviews – II
Chintan Panjwani1, Rashmi Thakur2

1Chintan Panjwani, M.E Computer Engineering Student, Thakur College of Engineering & Technology, Mumbai.
2Rashmi Thakur, Assistant Professor, Thakur College of Engineering & Technology, Mumbai 
Manuscript received on n September 06, 2019. | Revised Manuscript Received on n September 20, 2019. | Manuscript published on n September 20, 2019. | PP: 7-13 | Volume-5 Issue-6, September 2019. | Retrieval Number: D0928055419/2019©BEIESP | DOI: 10.35940/ijies.D0928.095619
Open Access | Ethics and Policies | Cite
© The Authors. Published By: Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (

Abstract: Enhnaced Dual Sentiment Analysis (EDSA) is an improved system which enhances the performance of the existing Dual Sentiment Analysis (DSA) which is implemented in literature. It mainly focuses on improving the efficiency of the existing system by making some modifications to the existing DSA approach. EDSA improves the classification accuracy of the public reviews. Apart from the classification accuracy other parameters considered in EDSA are precision, recall and fmeasure. In the first phase, a data pre-processing is performed to clean the data where subjectivity analysis is performed to obtain the subjective reviews and sentiment analysis is performed on subjective reviews only. Second phase deals with negation detection and sentiment word sreversal to obtain the reversed reviews. Third phase performs polarity calculation on the original and reversed reviews to obtain positive and negative reviews based on sentiment score of the reviews. Fourth phase performs the enhanced dual training and prediction where the positive and negative reviews are provided to various classifiers which provides the final results as the output. Final phase is the graphical representation of the various parameter values obtained from the previous phase which helps in comparing the results of the various classifiers.
Keywords: Bag of Words, Enhanced Dual Sentiment Analysis, Polarity Shift problem, Sentiment Analysis, Support Vector Machine