摘要
为降低由蜕变关系严密性较低引起的蜕变测试结果可靠性较低的问题,提出分类算法应用程序的蜕变测试方法。在对分类算法应用程序特征进行分析的基础上,分别从变换、增减两个方面构建了10条更加完善的蜕变关系,通过运用到蜕变测试的流程中,实现更高精度的分类结果测试。试验结果表明:测试方法可实现最大分类类别的平均相似度在96%左右,且整体的平均相似度可达到95%以上,具有较高的实际应用价值。
In order to reduce the low reliability of the transformation test results caused by the low rigor of the transformation relationship,a transformation test method for classification algorithm applications is proposed.Based on the analysis of the application features of classification algorithm,10 more perfect transformation relations are constructed from two aspects of transformation,increase and decrease,which are applied to the transformation test process to achieve higher accuracy of classification result test.The results show that the proposed test method can achieve the maximum average similarity of 96%,and the overall average similarity can reach more than 95%,which has high practical value.
作者
陈自力
CHEN Zi-li(School of Information and Intelligent Transportation,Fujian Chuanzheng Communications College,Fuzhou 350007,Fujian,China)
出处
《贵阳学院学报（自然科学版）》
2022年第1期1-4,共4页
Journal of Guiyang University：Natural Sciences
基金
福建省社会科学规划项目“基于案例驱动的事故灾害应急方案生成方法研究”(项目编号:FJ2019B079)
福建船政交通职业学院2018年科技项目“基于互联网+”的无车承运人监测系统的构建研究(项目编号:X19107001)。
关键词
分类算法
应用程序
蜕变关系
平均相似度
Classification algorithm
Application program
Metamorphosis relationship
Average similarity
