Chinese Proper Names Recognition Based On Pattern Matching

The recognition of proper names (PNs) is one of the tasks on Chinese natural language processing research, and is also the question that still have not been resolved properly. The building of large-scale corpus with high quality has been prior to all others in the area of natural language processing, since the analyses of corpus and automatic knowledge acquisition are growing high recognition. The main reason that influences the quality of corpus is that the proper names could not be recognized properly. The correct recognition of proper names must raise the segmentation quality of corpus. PNs recognition can provide support in the fields of natural language processing, such as information extraction, question answering, machine translation and so on.We proposed a pattern-matching-based method of proper names recognition aiming at three types of proper names (person name, location name, organization name) that appear most frequently. It can extract and classify proper names by combining searching and matching of outer-pattern and determinant of inner-pattern. The main works of this paper includes the below parts:1. Analyzing the structure of Chinese proper names, we built the set of inner-pattern of proper names (person name, location name, organization name);2. Analyzing the context of proper names that appears in text, we proposed outer-pattern of proper names, and automatic generation of PNs outer-pattern adopting clustering, evaluation etc.3. We proposed a method of competing categories of proper names recognition, combing multiple learners in parallel. It made proper names can be recognized with unified method. The proper names that have been recognized correctly can provide useful information for the recognition of other categories of PNs, and avoid wrong recognition of other categories,...