【摘要】 提出并实现了一个基于互信息的中文姓名识别方法.该方法充分挖掘姓名和其上下文信息的关联程度以及姓名用字之间关联程度的信息,引入互信息对其进行定量的描述;提出中文姓名的上下文互信息、内部互信息等概念,并对其建立了动态评价函数.开放测试结果表明,该方法有效地提高了中文姓名识别的效果,保证了较高的精确率和召回率.更多还原

【Abstract】 A method based on mutual information to identify Chinese names is proposed. The word association norms between Chinese names and their contexts as well as those among the characters or words in the names are studied, and then, mutual information is introduced to describe them quantitatively, the concepts of con-mutual information, inner-mutual information, etc. are also introduced. Lastly, the dynamic evaluation functions are built. The open tests on real data sets show that it is an effective method to improve the result of the identification, and high accuracy rate and recall rate are guaranteed.更多还原