Common Sense Knowledge Enhanced Opinion Mining for Social Media

Abstract

The vast quantity of available data and expansion of social networking sites has led to the  growth of opinion mining techniques. Nevertheless, opinion mining techniques, particularly  the traditional ones, are often challenged by the complex nature of communication on social  networks, which is contextual and versatile in nature. A new approach to opinion mining is  presented in this paper, which suggests using common sense knowledge in order to improve  the performance of sentiment analysis in social network environments. With the use of  extensive common sense knowledge networks and innovative natural language processing  capability, our method possesses a high degree of improvement in dealing with aspects such as  subtleties, sarcasm in social media texts, and unexpressed feelings in social media postings.  We provide a detailed experimental research framework including thorough methodologies,  results, and quantitative analysis which show that our technique works effectively on various  case studies across different social media sites and different topics.