构建软件测试领域不确定性知识图谱  

Construction of Knowledge Graph of Software Test Asset for Uncertain Knowledge

作　　者：罗浩榕 朱卫星 史涯晴 万进勇 LUO Hao-rong;ZHU Wei-xin;SHI Ya-qing;WAN Jin-yong(School of Command&Control Engineering,Army Engineering University of PLA,Nanjing 210001,China)

机构地区：[1]陆军工程大学指挥控制工程学院,江苏南京210001

出　　处：《计算机技术与发展》2022年第7期111-116,共6页Computer Technology and Development

基　　金：国家重点研发计划项目(2018YFB1403400)。

摘　　要：由于软件测试技术的不足和成本的限制,从历史软件测试资产中获取的可复用知识对象具有不确定性。为了更有效地复用历史测试知识,软件测试知识复用系统应当建模和评估知识的不确定性。模糊知识图谱是利用不确定性理论对典型知识图谱的一种扩展,在利用语义网络描述知识实体及其关系的基础上,通过模糊关系建模知识不确定性。通过分析软件测试资产,确定测试需求、测试用例和问题报告三类知识的表现形式和内涵,建立软件测试资产本体和领域特征本体作为知识图谱模式层,提出一种基于信息论的领域特征抽取方法,构建领域特征图谱。分析软件测试不确定性的产生原因,基于粗糙集理论对本体进行扩展,在软件测试资产和领域特征之间建立模糊关系。提出一种基于知识粒度的模糊置信度计算方法,用于构建软件测试模糊知识图谱,为软件测试中不确定性知识的复用提供了基础。通过一个指挥信息系统领域实例验证了方法的可行性。Due to the limitation of software testing technology and cost,the reusable knowledge objects obtained from historical software testing assets have uncertainties.To reuse historical testing knowledge more effectively,the software testing knowledge reuse system should model and evaluate these uncertainties.Fuzzy knowledge graph is an extension of typical knowledge graph based on uncertainty theory,which models knowledge uncertainty by fuzzy relationship on the basis of describing knowledge entities and their relationships by semantic network.By analyzing software test assets,the representations and content of test requirements,test cases and problem reports were determined,and the software test asset ontology and domain feature ontology were established as the pattern layer of knowledge graph.A domain feature extraction method based on information theory was proposed to construct domain feature graph.By analyzing the causes of uncertainty in software testing,the software test assets ontology is extended with rough set theory,and the fuzzy relationship between software testing assets and domain features is established.A fuzzy confidence calculation method based on knowledge granularity is proposed to construct fuzzy knowledge graph of software testing,which provides a basis for the reuse of uncertain knowledge in software testing.The feasibility of the method is verified by an example in the field of command information system.

关 键 词：软件测试 知识图谱 不确定性知识 粗糙集 知识粒度 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]