It is a very difficult problem in natural language processing to resolve the ambiguity of personal pronouns anaphora in a sentence or paragraph by computer according to semantic expression. Firstly, this paper focuses on finding out the personal names based on the maximal entropy model. Secondly, it categorizes Semantic Chunks and Sentence Category (SCs) based on the HNC theory. Thirdly, we chose 40 paragraphs in 2004 Athens Olympics as training corpus to make up the personal pronouns disambiguating rules. Finally, we chose another 40 paragraphs to exam the rules and processing steps by simulating computer's processing manually. We got a very high precision. Therefore, based on parallel semantic structure of HNC, the approach is effective for the disambiguating of Chinese personal pronouns.
展开▼
