Detecting Emotions in Social Affective Situations Using the EmotiNet Knowledge Base

The task of automatically detecting emotion in text is challenging. This is due to the fact that most of the times, textual expressions of affect are not direct - using emotion words - but result from the interpretation and assessment of the meaning of the concepts and their interaction, described in the chains of actions presented. This article presents the core of EmotiNet, a knowledge base (KB) for representing and storing affective reaction to real-life contexts and action chains described in text, and the methodology employed in designing, populating, extending and evaluating it. The basis of the design process is given by a set of self-reported affective situations in the International Survey on Emotion Antecedents and Reactions corpus. From the evaluation performed, we conclude that our final model represents a semantic resource appropriate for capturing and storing the semantics of real actions and predict the emotional responses triggered by chains of actions.