Emotion Detection in Textual Information by Semantic Role Labeling and Web Mining Techniques

Abstract

Automatic emotion detection in textual information is critical for the development of intelligent interfaces in many interactive multimedia applications. In the literature, existing approaches based on keyword spotting or statistic natural language process techniques, have limited success rate in free text emotion sensing applications. In this paper, we describe a system, developed in the framework of the National ChiNan University and LORIA collaboration, that associates semantic labeling and web mining techniques, to detect several basic emotions. A common sense knowledgebase – ConceptNet – is also used in order to retrieve some additional contextual information that can be used to retrieve appropriate background images for the presentation. Our objective is to adapt a multimedia presentation by detecting emotions contained in the textual information.