Emotion Recognition in Social Media: A Case Study About Tax Frauds

Abstract

Analyzing and understanding the relation of emotions and human computing interaction has become a necessity today. Indeed, sentiment analysis tools have gained special attention during the last years in order to facilitate and support the understanding and study of human affections. In this paper, we analyze an important Chilean tax fraud case by combining sentiment analysis and critical discourse analysis. We take as a case study, the tweets of the year 2018 that contain the #SQM hashtag. This case involves tax fraud and violations of political campaign laws. People from different political parties created fake invoices, which are then paid by SQM to be illegally used onto political parties violating campaign finance laws. Interesting results are obtained where we identify which topics and persons have a negative or positive connotation in the readers.