汉语词义排歧的另一种方法

Anothor Method of Chinese Word Sense Disambiguation Method

针对传统的基于义原同现频率的汉语词义排歧方法存在“盲目性”的不足,笔者根据《知网》中对概念定义的描述,分别计算多义词的每个义项与特征词的第一独立义原、其他独立义原、关系义原、符号义原之间的相关系数;最后通过比较多义词的每个义项与特征词之间的相关系数来决定多义词的义项.经过实验验证,该方法进一步提高了词义排歧的效果.

There is the fault of blindness in traditionary chinese word sense disambiguation method based on primitive co-occurrence data. On the ground of the depiction for the concept of Flownet,the author calculates these relation-moduluses between the first unattached primitive.the other unattached primitive.the relation primitive.the sign primitive of each sense of multivocal word and them of character-words. At last,the sense of muitivocal word is ehoosed by comparing these relation-moduluses between sense of multivoeal word and character-words. The result of experimentation showed that the effect of word sense disambiguation is improved.