Multi-emotion extraction from text based on linguistic analysis

Emotions as one of the important elements of human nature are a part of everyday communications of people. We can distinguish person's emotions from some outcome behaviors such as speech, facial expression, body movements and gestures. Another outcome behavior that reflects the inner states of the person is his/her grammar and written method. Because nowadays, people are more likely to use textual tools to make connection, emotion extraction from the text has attracted a lot of attention. This paper provides a framework for the extraction of emotions in the text. By considering that a text may contain more than one emotion that only one of them is text dominant emotion, our proposed method, has modeled emotion extraction problem as a multi-label classification problem by removing the fixed boundaries of emotions, and recognizes all the existing emotions in the sentence and also dominant emotion. The proposed method extracts emotions by using structural and semantic information in the sentence, linguistic information and machine learning techniques. The experiments have been done on multi-label dataset contains 629 sentences with eight emotional categories. Based on the results, our proposed method, compared with used multi-label learning methods (BR, RAKEL, MLANN) have shown a better performance.