基于数据驱动的软件可靠性预测模型  

Software Reliability Prediction Model Based on Data Driven

作　　者：农嘉[1] 罗峻文[1] NONG Jia, LUO Jun-wen(Department of Automobile and Information Engineering ,Guangxi Eco-engineering Vocational and Technical College, Liuzhou 545004, Guangxi, Chin)

机构地区：[1]广西生态工程职业技术学院汽车与信息工程系,广西柳州545007

出　　处：《内蒙古师范大学学报（自然科学汉文版）》2018年第1期49-54,共6页Journal of Inner Mongolia Normal University(Natural Science Edition)

基　　金：广西省教育厅基金资助项目(YB2014528)

摘　　要：当前以结果驱动为基础的预测方法存在预测收敛性差、准确性差等问题,为此提出基于数据驱动的软件可靠性预测模型.采集软件运行主机与控制服务器的数据,采用数据驱动的模型进行软件异常事件检测,构建软件演化模型;采用自适应持续性信息捕获算法进行软件可靠性建模,实现软件可靠性性约束参量的数据驱动,完成软件可靠性预测.仿真结果表明,采用该模型进行软件可靠性预测的准确性较高,预测过程的收敛性较好,从而提高了软件的寿命周期.In order to improve reliability and stability of the software,software reliability prediction is an important work.Current prediction method based on results driving has poor convergence and low accuracy,and a prediction model of software reliability based on data driven is proposed.Software running host and control server data are acquired,data driven model is used to detect software abnormal event to construct software evolution model,and adaptive continuous information acquisition algorithm is used for software reliability modeling to data drive for software reliability constraint parameter to complete software reliability prediction.The simulation results show that the accuracy of the software reliability prediction model is higher,the convergence of the prediction process is better,and the life cycle of the software is improved.

关 键 词：数据驱动 软件 可靠性预测 寿命周期 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]