摘要： 影响词义消歧的特征是多方面的,为考察上下文指示词这一消歧特征,提出了利用统计的方法抽取指示词用于词义消歧。实验表明,该方法是可行的,并为利用多特征进行汉语自动消歧提供了可靠的参考。

关键词: 自然语言处理, 词义消歧, 指示词, 相似性

Abstract: There are many different kinds of features contribute to the resolution of word meaning, in order to analyze the feature of word senseindicators, this paper presents a statistical method to extract indicators as semantic disambiguation. The experiments show that the strategy isworkable and provides credible reference for Chinese word sense disambiguation with multi-feature.

中图分类号: 

• TP391.1