QUANTIFIABLE APPROACH FOR EMOTION ANALYSIS OF TEXTUAL CORPUS

Abstract

To detect emotions of the writer in any textual corpus is a requirement that has been persistent in the industry. In the current phase of this research work, we are focusing only on the eight primary bipolar emotions suggested by Robert Plutchik[1] in his emotion wheel which are: anger/fear, joy/sadness, trust/disgust and anticipation/surprise. Now, the need for a textual corpus in which the instances labeled with their respective emotions, has been rising in the linguistic industry. Any corpus based Emotion Analysis approach belongs in one of the two categories: Keyword based approach and Ontology based approach. It was not hard to nd the textual corpus associated with the Ekman's six basic emotions but Plutchik[1] suggested thirty two emotions lying in a cone shaped 3D diagram, plus he also theo- rized the twenty four dyads which are mixed feelings on these eight primary emotions. That's why the manual annotation of textual corpus is required to study these complicated emo- tions. However, in the current phase of this research work, only the eight primary emotions are being considered. In the next phase, the dyads proposed by Plutchik will be observed and addition- ally the sixteen dyads proposed by Jessica Hagy on the advance level of Plutchik's emotion wheel will be considered either. Plus the role of these two parameters: "sensitivity" and "attention" can be observed to recognize the hidden and undetected emotions[2].