Abstract: This paper proposed a new method, for solving the problem of identifying important aspects from num erous online customer review .Determining important aspects form number of review increase the usability of unstructured review. There are number of existing methods available for solving this opinion mining problem. But, existing approaches failed in focusing on determining product aspects which are specifically commented or clearly mentioned in customer review. Therefore, we proposed new method for task of identifying important aspect based on the concept of intrinsic and extrinsic domain relevance score. In this paper, we first identify aspects and ,then calculate it’s intrinsic and extrinsic domain score. Aspects which are more relevant to the given domain yet not generic one tuned to be final aspects. Then, naive bayes classifier is used to determine opinion specified by reviewer on individual aspects Aspect ranking algorithm which is based on the concept of aspect frequency, opinion on aspect and relation between customer opinion on each aspect and it’s overall rating is used for ranking purpose. It is used to calculate individual aspect importance score and according to it’s importance score aspects are finally ranked.
Keywords: Product aspect, aspect ranking, sentiment classification, customer review, opinion mining, aspect identi fication, feature extraction, opinion feature.
Title: Aspect Ranking Based on Intrinsic and Extrinsic Domain Score
Author: Priti Sole, Mandar Kshirsagar
International Journal of Computer Science and Information Technology Research
ISSN 2348-1196 (print), ISSN 2348-120X (online)
Research Publish Journals
