User Interest Base Product Ranking Using Opinions’ Aspects

Abstract

Due to Online shopping trend, a vast number of reviews are generated on daily basis. These reviews help users in selection of products or services. Most of the existing works utilize these reviews in different research areas. Ranking products using these reviews is one of them. Recent studies in the area of research did not focus ranking products on the basis of user specification needs. Previous studies did not focus on ranking products on the basis of user specification. Most of the times the negative polarity of aspects given by users in their reviews is neglected while ranking the products. The main contribution of this study is to take negative polarity of the aspects into consideration, to improve the ranking task. To rank products on the basis of user specification, this research considered both polarities (either positive or negative) of the aspects. In which firstly review is broken down into the sentences to cover the cases in which a review contains multiple senti ments and various aspects. Thereafter aspects are extracted from the reviews of products given in the dataset and polarity of products is calculated on sentence level. Then the polarity is mapped with the aspects extracted from the specific sentence of the review. By using the weightage of the positive and negative polarities products are ranked ac cording to user specification needs. Experimental results show that users now possess the power change the weight of polarities and the resultant ranked list will be according to the user. It means if a user does not intend to include a specific polarity, he may ignore the same be assigning weight 0 to that. This study helps users in selection of good products or services. The instant work is performed for English language reviews. This research opens a gate to the natives of different countries to enhance the work for their native languages.