Sentiment Analysis or opinion mining is the process of getting decision support information from online reviews, information about any product, service event or person shared in blogs, forum and social networks. Sentiment analysis is extremely useful in social media monitoring as it allows us to gain an overview of the wider public opinion behind certain topics. The applications of sentiment analysis are broad and powerful.
Existing work handles the automatic keyword extraction through which representation of text document in a condensed way is possible. Existing studies makes an attempt to classify movie reviews using various supervised machine learning algorithms, such as Maximum Entropy (ME). The classification polarity of the document is limited only to text reviews and also it is observed that as the value of ’n’ in n-gram increases the classification accuracy decreases.
This project proposes the approach of integrating both the text and star ratings of user feedbacks using Naive Bayes (NB) and Maximum Entropy (ME) classification algorithms. It considers each term separately as single gram and multi-class classification has been implemented. In addition, interface is provided for reviewing the single user feedback. Hence, the accuracy of polarity of reviews is increased and the changes provided by star ratings are found.



maxentpickles folder has the list of trained models.
![Alt text](https://github.com/ash-sha/Opinion-Analysis/blob/master/UI1.png?raw=true)
![Alt text](https://github.com/ash-sha/Opinion-Analysis/blob/master/UI2.png?raw=true)
![Alt text](https://github.com/ash-sha/Opinion-Analysis/blob/master/UI3.png?raw=true)
![Alt text](https://github.com/ash-sha/Opinion-Analysis/blob/master/UI4.png?raw=true)