条件随机场模型在中文人名识别中的研究与实现

Research and Implementation of Conditional Random Field Model in Chinese Personal Name Recognition

中文人名的识别至今还是自然语言研究领域一个比较困难的课题,因此提出一种基于条件随机场模型的文中人名识别方法。条件随机场模型是一种无向图模型,有效避免有向图在标记的过程中出现偏执的问题,并且通过二次识别,有效解决人名在上下文环境中的识别问题。通过实验分析,基于条件随机场模型的人名识别能比较准确地识别出中文的人名。

In the field of natural language research, recognition of Chinese name is still a difficult issue, so points out a kind of recognition of Chinese name based on conditional random fields model. Conditional random fields model which is a kind of undirected graph models that can avoid the bias problems which digraph produces in the process of marking. By secondary recognition, it can solve these issues of recognizing name on the context. Through the analysis of experiments, name recognition based on conditional random fields is able to recognize Chinese name more accurate.