Automatic detection of emotions in Twitter data
Jaishree Ranganathan, Nikhil Hedge, Allen S. Irudayaraj, Angelina A. Tzacheva · 2018
Social media data is one of the promising datasets to mine meaningful insights with applications in business and social science. Emotion mining has significant importance in the field of psychology, cognitive science, and linguistics etc. Recently, textual emotion mining has gained attraction in modern science applications. In this paper, we propose an approach which builds a corpus of tweets and related fields where each tweet is classified with respective emotion based on lexicon, and emoticons. Also, we have developed decision tree classifier, decision forest, and rule-based classifier for automatic classification of emotion based on the labeled corpus. The method is implemented in Apache Spark for scalability and BigData accommodation. Results show higher classification accuracy than previous works.