Determining Emotional Profile Based on Microblogging Analysis

Abstract

In general, groups of people are formed because of the similarities and affinities that members have with each other. Musical preferences, soccer teams or even similar behaviours are examples of similarities and affinities that motivate group formation. In social media, identifying these affinities is a difficult task because personal information is not easily identified. In this paper we present an alternative to identifying similarities between authors and their most frequent audience in Twitter, using emotional and grammatical writing style analysis. Through this study it is possible to define the creation of an emotional profile entirely based on the interactions of people, thus allowing software like chatbots to “learn emotions” and provide emotionally acceptable responses.