基于支撑向量机的软件测试方法

Software Testing Method Based on Support Vector Machine

当软件运行结果为海量的非量化信息时,传统测试方法无法进行有效测试。引入基于支撑向量机(SVM)测试模块,利用适当的核模型对灰度、图形及态势结果等非量化信息进行测试,并通过构建MATLAB环境下软件测试框架,实现对海量数据的有效测试,从而提高测试质量与效率。

When the running results of a software are massive and non-quantizing information,the traditional testing method cannot be effectively tested.A testing module based on support vector machine（SVM）is introduced.By selecting appropriate kernel model,non-quantitative information,such as color,graphics and situation result is tested.Through building a software testing framework in MATLAB environment,the massive data can be tested,thus the testing quality and the testing efficiency are improved.