OPINE_NEG: AN APPROACH TO DETECTING NEGATIONS AND INTENSIFIERS USING SOCIAL MEDIA DATA

Abstract: With the abundance of online information available today, there are numerous opportunities and challenges that arise for both consumers and internet users alike. Managing customer satisfaction is a critical performance indicator for running a successful business. Opinion Mining, also called as sentiment analysis, is the method of digitally assessing public feelings, emotions, opinions, and services. This text mining technique and natural language processing are used to automatically extract, classify, and summarize sentiments and feelings expressed in online text. This study focuses on subjective terms in the user opinions collected from microblogging sites and analyses the semantic orientation of reviews to identify and classify negations, intensifiers, slang words, and emoticons. This study proposes an improved lexicon- based dictionary approach using a rule-based classification scheme to enhance the accuracy of opinion mining on Big Data. To test the efficiency of the research work, user reviews in four domains - mobile, health, electronic, and vehicle - were analysed using Twitter datasets.

Keywords: Opinion Mining, Microblog data, Negations, Intensifiers, Big Data.