Fuzzy Based Summarization of Product Reviews for Better Analysis

Background/Objectives: There is a tremendous growth in the online marketing where customers buy a product and leave a comment on it about their experience. These experiences which are in the form of Reviews help in two ways. Methods/Statistical Analysis: Firstly, the buyer will have a clear idea about the product pros and cons. Secondly; manufacturer will also find them helpful to make the user experience better by improving the product or service in negative areas. This converges at a point where, if the user reviews are thousands in number for a single product, we can propose a system which provides the summary of all user generated reviews. This is what motivated opinion mining systems to summarize the user review. Opinion mining is the current technology which can classify the review documents to summarize them. Findings: This paper implements the opinion mining based on fuzzy logic to improve classification of reviews for generating the concise summary about the product. Application/Improvements: This is a Feature based sentiment classification which is a multistep process which involves pre-processing phase, fuzzy score to classify each review, training the Naive bayes classifier, evaluating each sentence in the test set depending on the trained classifier and ranking the sentences for each feature. Thus sentences evaluated are a fine grained classification to better summarize the reviews.