Graph-Based Sentiment Analysis Model for E-Commerce Websites' Data

摘要

E-Commerce has evolved tremendously in the past few years. To enhance the existing business position, the commercial sites need to understand the underlying sentiment of the customers. To do so, efficient sentiment analysis technique is highly desirable in order to deeply understand the underlying meaning and sentiment of the customers. This paper proposes an effective sentiment analysis model that makes use of graph-based keyword extraction using degree centrality measure and domain dedicated polarity assignment techniques for the advanced analysis of mobile handset reviews collected from different electronic commercial sites. The proposed model outperforms some of the existing models.