词义消歧是自然语言处理中的一个关键问题,为提高大规模词义消歧的准确率,提出了一种基于模板的无导词义消歧方法.利用多义词不同义项的同义或近义单义词对该义项进行表述,综合考虑共现词出现的位置、上下文距离及出现频次,据此构造语境模板,有效地解决了多义词义项确定的困难.实验结果表明,本文提出的方法在消歧性能方面有较明显的改善.%Word sense disambiguation is a key problem in NLP. This essay proposes a template-based unsupervised word sense disambiguation method to improve the precision. We describe polysemous word by the synonymies of different sense and construct the context template considering position, context distance and frequency of the co-occurrence term. Experiment shows this method could improve the performance of word sense disambiguation.
展开▼