大规模文本计算机音字相互转换技术的研究
A COMPUTER SYLLABLE TO CHARACTER CONVERSION TECHNIQUE BASED ON LARGE SCALE CORPUS
-
摘要: 文中介绍了一种利用马尔可夫统计文法和自然语言理解技术相结合的计算语言模型进行汉语音节到汉字的机器自动相互转换技术.该技术集成了统计文法和句法分析方法各自的优点,可以实用于计算机大规模真实文本的音字相互转换,并且具有处理语言深度递归现象和远距离约束关系的能力.其音字转换算法已被应用到大词表非特定人孤立词语音识别系统和中文WIN-DOWS95、WINDOWSNT4.0的智能语句输入法上.Abstract: A Chinese syllable to character conversion system is described,which is based on a rule statistic hybrid language model.A good performance is obtained in processing Chinese language recursive nature and long distance constrained relations.The system has been applied to the Chinese keyboard intelligent input method on Chinese Windows 95,Windows NT4.0,and the isolated word speech recognition system.
下载: