An Exploration of Features for Recognizing Word Emotion

Emotion words have been well used as the most obvious choice as feature in the task of textual emotion recognition and automatic emotion lexicon construction. In this work, we explore features for recognizing word emotion. Based on RenCECps (an annotated emotion corpus) and MaxEnt (Maximum entropy) model, several contextual features and their combination have been experimented. Then PLSA (probabilistic latent semantic analysis) is used to get semantic feature by clustering words and sentences. The experimental results demonstrate the effectiveness of using semantic feature for word emotion recognition. After that, "word emotion components" is proposed to describe the combined basic emotions in a word. A significant performance improvement over contextual and semantic features was observed after adding word emotion components as feature.

• Type: Journal
• Year: 2010
• Where: COLING
• Authors: Changqin Quan, Fuji Ren
• Keywords: COLING 2010, Computational Linguistics, Emotion, Word Emotion, Word Emotion Components