Detecting emotions in Greek sentences: towards an open-source affect-awareness framework

Περίληψη

Considering the highly important role of automated emotion recognition in the development of affective intelligent systems, this study investigates a real-time sentence-level emotion recognition method. More specifically, this paper describes the construction process of an open-source library that can be used for the detection of emotions from sentences comprising Greek or English words. The proposed library was developed, based on the Synesketch software library, utilizing Ekman’s basic emotion theory. Greek sentences are classified to six emotional types defined by Ekman (happiness, sadness, anger, fear, disgust and surprise).