基于知网的软件故障案例语义表示方法

Semantic Representation of Software Fault Case Based on How-Net

摘要:

为解决采用静态框架描述故障案例存在的问题,提出一种基于知网的软件故障案例语义表示方法。通过分析案例的语义匹配准则,给出表示方法的具体语法描述规范。在此基础上,结合实际案例,分析软件故障领域中实体、事件、属性和属性值4类义原及其相互关系,并用KDML语言进行统一描述。实验表明,与传统基于关键字的模糊匹配方法相比,采用基于知网的匹配方法能够有效提高CBR系统中故障案例检索的准确率。

Abstract:

In order to solve fuzzy semantic matching problem of the software fault case, a kind of semantic knowledge representation was introduced in this paper. By analyzing semantic matching criteria among cases, the grammar of representation was defined first. Further, a consistent definition used by KDML was adopted to explain the semantic atom of entity, event, attribute and value of attribute, as well as relationship between them. The experiment showed that this method had more accurate retrieving precision than the method based on Keywords matching.