摘要
针对软件测试领域测试用例复用性差、涉及学科专业复杂的特点,提出了基于深度学习的软件测试知识获取及应用方法,分别构建了软件测试知识模型和软件测试用例主题模型,并给出了软件测试领域知识论元识别的流程和方法。根据案例验证结果,基于建立的知识模型进行软件测试用例搜索时,模型加强了搜索的语义理解能力,并且使用基于测试用例建立的语义向量空间后,测试用例的复用正确率有了明显的上升,表明了基于深度学习可以有效建立软件测试知识库,为软件测试用例的复用和智能化编辑提供参考。
For the characteristics of poor reusability of test cases and involving multiple disciplines in the field of software testing,a deep learning-based software testing knowledge acquisition and application method is proposed.The software testing knowledge model and the software testing case topic model are respectively constructed,and the process and method of knowledge argument identification in the field of software testing are given.Experiments show that searching using only keywords sometimes results in no reuse of test cases.After semantic understanding,the search results have improved significantly.After using the semantic vector space based on test cases,the correct rate of test case reuse has increased significantly.Based on the deep learning method,the software testing knowledge base can be effectively established,which provides ideas for the reuse and intelligent editing of software test cases.
作者
李超
张翔
彭珲
崔龙飞
LI Chao;ZHANG Xiang;PENG Hui;CUI Longfei(63892 Troops of PLA,Luoyang Henan 471003)
出处
《软件》
2023年第5期61-66,共6页
Software
关键词
深度学习
软件测试
知识获取
模型
神经网络
deep learning
software testing
knowledge acquisition
models
neural networks
作者简介
李超(1986-),男,河南信阳人,硕士研究生,助理研究员,研究方向:软件测评。
