Date of Publication :7th April 2017

Abstract: Over the last few years, opinions are voiced in the form of blogs, tweets, social media and review over the web. These opinions do matter, as, we always consider other persons view in our decision making process. Proposed fuzzy technique finds out how positively or negatively the masses have responded to the launched product, based on voiced opinions. Using a supervised rule based technique, fuzzy values are assigned to opinions, to know the measure of its positivity or negativity. Steps involve, first mining the opinion about the product from ecommerce sites. Then extraction of the opinion phrases as per pre-defined phrase patterns. Next, assigning of fuzzy weights to the opinion words using SentiWordNet and measuring the strength of opinion phrases. And, finally summarization of people’s views to yield product recommendation. Proposed fuzzy technique is an enhancement over the existing opinion mining techniques which does binary polarity determination.

Reference :