Multi-layered sentiment analytical model for product review mining

摘要

The trend of the online shopping has grown over the recent years, which is clearly visible from the profit statistics posted by the big e-commerce companies in the recent years such as flipkart, snapdeal, etc. The online shopping brutally skips the touch and feel of the products which we generally evaluate as the primary factor during purchasing the commodities from the shops or stores. Assessing the quality of the product also becomes very important, which can be generally discovered by applying the opinion mining techniques over the review data. The product review classification is the mechanism which is used to analyze the sentiment or opinion in the reviews posted by the users in order to prepare the product report. In this paper, the mechanism has been proposed for the opinion mining over the text review data for the generation of the product review report, which is based upon the multiple features altogether. The product report can reveal the several positive and negative points about the specific product, which can play the significant role in the selection of the products on the online portals. The proposed model has been evaluated on the basis of the performance parameters of the precision, recall and polarity based accuracy assessment, which gives the overall perspective about the overall accuracy. The proposed model has been clearly defined better than the existing models, when assessed on the basis of the given parameters.