基于知网的关系网络的构建

为了更好地利用知网中的语义关系,提出了一种将知网中的各种隐含关系形式化表示的方法.该方法通过仔细分析知网中关系标知符、动态角色以及例句所描述的各种语义关系,以概念、特征、关系为基础,构建了一基于知网的关系网络.利用该关系网络,既可以高效地查找到词语的概念与概念之间的语义关系,也可以查找到概念的属性之间的语义关系.给出了该方法的实现算法,并用实例说明了该方法的优点.

Construction of relation network based on Hownet

To fully utilize semantic relations in Hownet,a method to capture implicit semantic relations in Hownet was introlduced.By analyzing relations described by pointers,event roles and examples,a so called relation network was constructed based on Hownet's concepts,features and relations.The relation network can be used to determine various relations efficiently between concepts and features.An algorithm was presented to realize the method.The merits of the method was illuminated by example.