摘要
介绍了一种利用《同义词词林》和训练语料生成义类代码同现频率矩阵 ,以此作为资源进行真实语料中多义词的词义排歧。由于该方法采用无指导的学习方法 ,可以免除人工标注的开支 。
Word sense disambiguation is one of the difficult problems and a key point in NLP.in this paper, an unsupervised method is discribed. Based on corpus and a dictionary on word defintition synonyms, the word cooccirenced data matrix can be obtained, and using the matrix to realize the word disambiguation, this method is proved feasible.
出处
《河南职业技术师范学院学报》
2002年第1期53-54,57,共3页
Journal of Henan Vocation-Technical Teachers College