基于自适应高斯混合模型的软件测试用例集约简算法研究

Research on Software Testing Case Reduction Algorithm Based on Adaptive Gaussian Mixture Model

摘要:

为了最大化的找出软件测试用例集中的相似用例,实现对用例的最优精简,提出了一种自适应的高斯混合模型。提出的模型使用K-means初始化EM,自适应地确定聚类簇数目,在此过程中能够评判聚类结果,同时给出式高斯混合模型的所有参数,这些参数作为各个聚类簇进行新一轮迭代计算的参数,最终得到的结果更趋于最优解。实验结果表明,相对现有的高斯混合模型和模糊K-Means聚类模型等算法,本文提出的自适应高斯混合模型算法能够最小化软件测试用例集,约简后的用例所覆盖的范围相对更广,测试出的软件错误率较高,对软件测试用例集多变的适应性好。

Abstract:

In order to find out the similar test cases in the software test case set and realize the optimal simplification of test cases, an adaptive Gaussian mixture model is proposed. The proposed model uses K-means to initialize EM, adaptively determines the number of clusters. In this process, the clustering results can be evaluated. At the same time, all the parameters of the Gaussian mixture model are given. These parameters are used as the parameters of each cluster for a new round of iterative calculation, and the final results tend to be the optimal solution. The experimental results show that, compared with the existing Gaussian mixture model and fuzzy K-means clustering model, the adaptive Gaussian mixture model algorithm proposed in this paper can minimize the software test case set, and the reduced cases cover a wider range. The software error rate is higher, and it has good adaptability to the variety of software test case set.