语言模型中一种改进的最大熵方法及其应用

An Improved Maximum Entropy Language Model and Its Application

最大熵方法是建立统计语言模型的一种有效的方法，具有较强的知识表达能力．但是，在用现有的最大熵方法建立统计模型时存在计算量大的问题．针对这一问题，提出了一种改进的最大熵方法．该方法使用互信息的概念，通过Ｚ－测试进行特征选择．将该方法应用于汉语的义项排歧中，实验表明。

The maximum entropy approach is proved to be expressive and effective for the statistics language modeling, but it suffers from the computational expensiveness of the model building. An improved maximum entropy approach which makes use of mutual information of information theory to select features based on Ztest is proposed. The approach is applied to Chinese word sense disambiguation. The experiments show that it has higher efficiency and precision.