A Multimodal Approach To Detect User's Emotion☆

Abstract Emotion is a feeling, a behaviour, a state of mind of a person that is caused due to internal or external factors. Among the numerous emotions, we have considered a few emotions like happy, sad, fear, anger, surprise, love, hate, exhaust and disgust in the paper. Emotions are detected predominantly using analysis of textual contents even though there are other means such as analysis of speech, facial expressions etc. In this paper, we present a multimodal approach to detect emotions by analysing elements like the dynamics of the user's keystroke, semantics of the text and variations of the heart rate. Results of our experiments have shown that consideration of these emotion elements is found to be effective in detecting user's emotions.