【摘要】 本文根据100万字的科技资料中出现3129个不同汉字的字频统计数据,将汉字按字频大小排序,得出字序n较小时,字频近似于Zipt分布,大n时趋于指数分布的规律。根据这种分布,得到汉字的一维熵和汉字最佳编码的平均码长。又根据汉语拼音统计数据,估计汉字多维熵以及汉语拼音熵,并据以分析汉语拼音输入方案。更多还原

【Abstract】 Serial numbers are given to 3129 different Chinese characters according to their frequency of occurrence in chinese scientific publications amounting to 1 million characters. The result is, with small serial number n, the frequency of occurrence is approximately Zipf distribution, while with big n the distribution approaches exponential. The frist order entropy and the average code length of the optimum code of the Chinese characters are evaluated. The multiple order entropy and the phonetic entropy of the chinese characters are also evaluated, and an input scheme is analyzed.更多还原