摘要
针对软件缺陷测试任务中的准确度问题,提出一种基于优化BP神经网路的软件缺陷预测方法 .该方法首先采用4层BP神经网络构建多层感知模型,并结合模糊控制原理实现任意复杂非线性关系逼近.然后通过灰狼优化算法克服BP神经网络的局部搜索陷入,从而解决其参数设置依赖性问题.实验结果表明,相比于PSO-BP算法和SA-BP算法,该算法的仿真拟合效果最优,表现出了更高的软件缺陷预测准确度.
Aiming at the accuracy problem in software defect testing task,a software defect prediction method based on optimized BP neural network is proposed.First,the 4 level BP neural network is used to build the multilayer perception model,and the fuzzy control theory is applied to achieve any complex nonlinear relation approximation.Then the Grey Wolf optimization algorithm is used to overcome the local search trapped in BP neural network,so as to solve the problem of parameter setting dependency.The experimental results show that compared with the PSO-BP algorithm and the SA-BP algorithm,the simulation results of the proposed algorithm are optimal,showing a higher accuracy of software defect prediction.
作者
曾毅
张福泉
ZENG Yi;ZHANG Fu- quan(Xingjian College of Science and Liberal Arts,Guangxi University,Nanning 530005;School of Software,Beijing Institute of Technology,Beijing 100081 China)
出处
《湘潭大学自然科学学报》
CAS
2018年第2期100-103,共4页
Natural Science Journal of Xiangtan University
基金
福建省科技厅省引导性项目(2018H0028)
文化部国家科技支撑计划项目(2012BAH38F00)
关键词
软件测试
缺陷预测
准确度
BP神经网络
灰狼算法
软件可靠性
software testing
defect prediction
accuracy
BP neural network
Grey Wolf algorithm
software reliability
