TEXTUAL EMOTION RECOGNITION FOR ENHANCING SOCIAL PRESENCE IN ONLINE COMMUNICATIONS

Abstract

This paper presents an emotion recognition model that assesses the affective content from textual messages. The focus is on emotion recognition from online non verbal textual symbols of vocalics (e.g., the use of capitals and use of punctuation “!” and “!!s!” or “?” and “???”, length of response, etc), and those of chronemics (e.g. time to respond to an email or to a discussion posting, the length of the response, etc) that are used in text. The model employs a supervised machine learning approach to recognize six basic emotions (anger, disgust, fear, happiness, sadness and surprise) using an emotion-annotated training set composed of online chat messages exchanged by university students. Naive Bayes algorithm is used to classify messages to the mentioned emotional classes based on a variety of features. The model also takes into consideration the evolving language particularly the language used in online chat where people tend to use an informal style of writing. Observations from informal experiments comparing a chat system integrated with the emotion estimation model with a conventional chat system suggest that an online interface that conveys emotional information helps online users to interact with each other more efficiently thus providing an enhanced social presence.