Research On Named Entity Recognition Based On Rules

Chinese word segmentation is the first step in natural language processing. In practice, Chinese word segmentation subject to many constraints, unknown word is one of the important factors impact the accuracy. Unknown words mainly contains person's name, place name, organization name and other named entity. Therefore, integrate named entity recognition into the process of Chinese word segmentation plays an important role in improving the accuracy of Chinese word segmentation. In addition, the named entity recognition research has important theoretical significance and practical value for information extraction, information retrieval, machine translation, text classification applications, the realization.Contributions of the dissertation are as follows:(1) The model of integrating named entity recognition into the process of Chinese word segmentation is proposed, Recognize Named Entity in the process of Chinese word segmentation. This method can reduced the Chinese lexical segmentation error caused by he unknown words such as named entity and enhance the accuracy of Chinese word segmentation.(2) Chinese name classification system is Constructed based on Ontology. Through this method, Chinese name knowledge will be divided into several levels. Low-level domain knowledge is the basis of high-level and high-level domain knowledge is the generalization and summary of low-level. This approach greatly improves the maintainability of the Chiese names knowledge.(3) The rule system for named entity recognition is built. Then use rule matching method to recognize named entity. This recognition system has an ability of self-learning. Recognize named entity at the same time analyze the results to produce new rules and add them to rule system. Experiments show that through this metod we can get good results of named entity recognition.