摘要
汉语中词的兼类是一个普遍存在的现象。任何工程化的汉语句法分析系统都不能回避这个重要而难以解决的歧义问题。本文根据汉英机器翻译系统CEMT—Ⅲ的有2万词条的机器词典进行了统计,其中兼类词占7.7％,刪CEMT—Ⅲ系统采用多级渐进处理策略,将确定性推理和非确定性推理相结合,实现了汉语词的兼类自动消除机制。
Category ambiguity in Chinese is an universal phenomenon. The percentage of ambiguous category words is 7.7% in the machine dictionary with 20,000 items of CEMT-Ⅲ Chinese-English Machine Translation System. A practical Chinese parsing system can never ignore this important ambiguous problem. CEMT-III system has adopted a stategy of gradually-processing, and combined defined inference and non-defined inference. A full automatic mechanism for category ambiguity solving has been accomplished.
出处
《中文信息学报》
CSCD
1993年第4期52-59,共8页
Journal of Chinese Information Processing
