Research Of Key Issues In Coreference Resolution Of Chinese

To avoid repetition in natural language, people always use pronoun, alias and abbreviation to refer a given entity. And this phenomenon is called coreference. As an important research area of Natural Language Processing (NLP), the task of coreference resolution is to find these coreference phenomenon.This dissertation focuses on key issues of Chinese coreference resolution and presents a Chinese noun phrase coreference resolution system that based on the supervised learning approach and the unsupervised clustering approach respectively. The major contribution of this dissertation is as follow:1. It analyzes the difference between the Chinese and English noun phrase coreference resolution. The dissertation also discusses the issues in Chinese noun phrase coreference resolution.2. According to the task of Chinese noun phrase coreference resolution, it introduces some effective features to construct a Chinese coreference resolution system based on the supervised approach. It also discusses the contribution of each feature and their combination.3. It proposes an unsupervised Chinese noun phrase coreference resolution platform based on some effect features and incompatibility functions. This platform can run on a small size corpus and reduce the influence of corpus to the Chinese noun phrase coreference resolution.4. It discusses the effect of the quantity and the quality of the corpus on the Chinese noun phrase coreference resolution and discusses the relation between different types of noun phrases and the quantity of the corpus on the Chinese noun phrase coreference resolution.