With the emergence of Affective Computing in the late nineties (Picard, 1997), several researchers in different computer science areas, e.g., Natural Language Processing (NLP), Human Computer Interaction (HCI), Intelligent Tutoring Systems (ITS), etc. are interested more and more in emotions. Their aim is to develop machines that can detect users' emotions and express different kinds of emotion. The most natural way for a computer to automatic emotion recognition of the user is to detect his emotional state from the text that he entered in a blog, an online chat site, or in another form of text.
In this talk, I will present some existing approaches for automatic emotion recognition. I will focus particularly on analysis of emotion in text. To this end, I adopted a supervised machine learning approach to recognize six basic emotions (anger, disgust, fear, happiness, sadness and surprise) using a heterogeneous emotion-annotated dataset which combines news headlines, fairy tales and blogs. Different features sets such as, lexical emotion features from WordNetAffect, bags of words, and N-grams were used.