Abstract: Individuals as well as institutions are paying increasing attention to sentiment analysis. Companies are interested in what bloggers are saying about their products. Customers are interested in knowing the reviews about the products/services which fall in the same category. It becomes essential for them to know and hence decide which of the products/services suit them the best.
Our application performs the task of classifying a given piece of natural language text (be it a short remark, blog post, or full-blown product review) not according to its topic (as in standard text classification) but according to the opinions expressed in it. It gathers reviews and blogs for the desired service query, analyses the sentiment/ opinion behind it, summarizes them and ultimately presents it to the user. This decision provided by the application thus aids the customer in making the appropriate choice. The user is segregated into categories, which gives him the privilege of considering the reviews of like-minded people. This hence improves the quality of the result which now becomes reliable. Reviews about services like wireless internet services, DTH service, insurance policies, loan providers and many more can be evaluated. The user can then make the desired use of the application for decision making.
We aim to gather the reviews through the feed mode or crawling mode. The sentence is then analyzed based on the combination of adjectives, verbs and adverbs to conclude the polarity of it. The strength of the sentence/paragraph can further be represented on a continuous scale of -1(maximally negative) to +1(maximally positive). A survey can be conducted to compare the efficiency of the application with manual sentiment analysis of the review.
The project is potential for many extensions which can be incorporated in the application. For eg. a DSS(decision support system) can be included to provide a more refined and a personalized summary. Also the project can be extended for multilingual systems, which makes it usable for non-English population also. Also, the project can be used as a back-end application for companies and institutions as a part of their CRM. The company can hence utilize the result for improvising the quality of their products/services.