汉字最高阶条件熵及其实验测定

在文字的信息熵中，最高阶条件熵的最本质的，最重要的，是度量文字信息量的主要依据。本文论述了汉字最高阶条件熵的基本内涵和数学模式，以及最高阶条件熵与句熵的内在联系，结合汉字在语言中的实际使用情况，提出了在汉语句子的基础上，通过可控的，顺序猜字试验，求出试验句子的句熵与对应的最高阶条件熵的方法，并建立了为之服务的计算机测试系统，初步试验证明，机测系统是成功的，取得的数据是可靠的。

Highest Order Entropy of Chinese Characters and its Evaluation in Experiments

Information and entropy are two very important foundamental concepts in communication theory and linguistic engineering. The highest order conditional entropy of character is of particular significance in comparison with other entropies of characters in the same language. Based on an analysis of the uncertainty of Chinese characters in different situations,and taking into account the entropies of single Chinese characters (may be measured in logPi), this paper proposes a method to be used in sequence-guessing of Chinese sentences.Thus, an experimental mode of sequence-guessing in Chinese characters and a computer testing system are successfully established. Some experiments on Chinese sentences have been done on the computer, and the obatined data are reliable. It is clear that our entire experiment procedure of senquence-guessing accords fully with the essential features of highest order entropies. Therefore,it will not be difficult to obtain a reasonable value rang in the highest order entropies of Chinese characters by enlarging the sequence-guessing experiments.