A novel formal computational model of Chinese text sentiment based on corpus

To give consideration to both human intuitive understanding and the information processing requirement of text sentiment, this paper describes a new formal computational model of Chinese text sentiment. The model includes two parts: formal presentation and formal computing. The former is based on seven-basic sentiment categories, namely “happy”, “anger”, “sadness”, “fear”, “love”, “hate” and “desire”, and every sentiment can be formally expressed by a seven-dimension vector. The latter analyses the relationship of the seven basic sentiment categories and then makes mathematical orthogonal based on the relationship in order to support the vector operation of sentiment computing. Quantifying processes of the given model based on corpus are detailed. An example shows that the model has capabilities of semantic interpretation and calculation, which can be the basic formal statement for information processing of text sentiment.