基于层次聚类的中文人名消歧方法研究

Chinese Personal Name Disambiguation:Based on Hierarchical Agglomerative Clustering Approach

人名消歧近来成为自然语言处理中的一个热点问题。由于中文的复杂性,中文人名消歧被认为比英文人名消歧更困难。本文在使用层次凝聚聚类算法的基础上,主要探讨了中文人名的识别对中文人名消歧的影响以及中文人名消歧有效特征的自动提取。实验证明,特征融合是提高系统性能的有效方法。中国中文信息学会与SIGHAN组织的评测表明本文所提出的方法是有效的。

Personal name disambiguation has became a active area in the natural language processing. Because of the complexity of the Chinese, the Chinese personal name disambiguation is considered more difficult than the English personal name disambiguation. Using hierarchical agglomerative clustering,this paper focuses on the impact of Chinese personal name detection on Chinese personal name disambiguation, and the automatic extraction of effective feature. The experimental results show that the proposed feature integration method could improve the system’s performance significantly. The result in the evaluation jointly organized by Chinese Information Processing Society of China (CIPS) and SIGHAN show that our method is promising.