基于语义相似度计算的词汇语义自动分类系统

词汇语义分类在文本聚类、信息检索、机器翻译等多个研究领域中拥有重要的理论及实践意义.介绍的知网语义相似度计算的词汇语义自动分类系统通过设计双向索引结构.高效的组织和挖掘了知网已有数据资源,并利用成熟的知网词语相似度计算方法,为词汇语义自动分类提供了不同于统计方法的新思路.目前系统的研究已取得实质性成果.在知网义原1564个分类的基础上,对知网提供的6万余条汉语常用词进行初步语义分类,进而开发了二次分类模块,针对初步分类结果进一步细化为适合实际需要的子类.实验结果证明该系统在分类性能上明显优于基于统计方法的分类系统,所作分类更加细腻、平滑.

A Word Semantic Automatic Classification System Based on Word Similarity Computation

Word Semantic Classification has its theoretical and practical significance in various research fields,such as text classification,information retrieval and mechanical translation.Through designing a two-way retrieval structure,this paper proposes a HowNet-based word semantic automatic classification system that has efficiently organized and excavated data resources in HowNet,and that has utilized HowNet's word similarity computation method,thus shedding a new light in word semantic automatic classification...