Author : Amruta Sankhe and Prachi Gharpure

The rapid growth of the Web in the last decade makes it the largest publicly accessible data source in the world. There are several websites which allows user to post reviews on particular product or service. Customers also want to know the opinions of existing users before they use a service or purchase a product; even businesses want to find public or consumer opinions about their products and services. Hence there is need to do automatic feature based opinion mining for buyers to take smart decision. However, since the polarity of features is varied according to domain the task of feature based summary is challenging itself.This paper presents an approach for mining online user reviews to generate feature-based sentiment analysis that can guide a user in making an online purchase. In this work, feature based sentiment analysis on car domain is presented. In which features are extracted using domain ontology, then by using the feature description table the features are classified into positive and negative polarity. Since the polarity of each feature is varied according to the domain. The outcome of the system is a set of reviews organized by their degree of positivity and negativity based on each feature. This system helps to reduce the manual effort of evaluating reviews according to features in which user is interested. The polarity obtained for each feature from our approach is with good average accuracy.

Keywords: opinion mining, sentiment analysis, feature-based opinion mining, summarization, semantic web.

Article published in International Journal of Current  Engineering  and Technology, Vol.4,No.2 (April- 2014)