中文领域本体概念层次获取方法对比研究

Contrast research of Chinese domain ontology concept hierarchy induction methods

概念层次是本体的基本骨架,而概念层次的获取又是本体学习中非常重要的一步。国际上对于概念层次获取的研究绝大部分都集中于英语,国内在该方面的研究还处于起步阶段,而目前已有的处理英文的方法用于处理中文效果如何,在国内还没有这方面的报道。重点比较了能用于获取中文概念层次的方法,并详细分析了各种参数的不同取值对结果的影响。结果表明在相同实验语料背景下,层次聚类法中基于VSM的方法效果最好。

Concept hierarchy is the basic component of ontology. And the concept hierarchy induction plays a very important role in ontology learning. Most researches of concept hierarchy induction focus on English corpus aboard, but which is at the beginning stage at home. There is no-discussion about the performance of methods used in english corpus to deal with Chinese corpus. This paper compared the concept hierarchy induction methods which could deal with Chinese corpus and analyzed how the variable values of parameter impact the result. Experimental results show hierarchical clustering based on VSM method achieves better performance than other methods.