马振宇, 吴纬, 张威, 刘福胜, 韩坤. 相关系数对软件可靠性验证测试的影响分析[J]. 微电子学与计算机, 2018, 35(5): 69-73.
引用本文: 马振宇, 吴纬, 张威, 刘福胜, 韩坤. 相关系数对软件可靠性验证测试的影响分析[J]. 微电子学与计算机, 2018, 35(5): 69-73.
MA Zhen-yu, WU Wei, ZHANG Wei, LIU Fu-sheng, HAN Kun. Analysis of Correlation Coefficient Influence on Software Reliability Demonstration Testing[J]. Microelectronics & Computer, 2018, 35(5): 69-73.
Citation: MA Zhen-yu, WU Wei, ZHANG Wei, LIU Fu-sheng, HAN Kun. Analysis of Correlation Coefficient Influence on Software Reliability Demonstration Testing[J]. Microelectronics & Computer, 2018, 35(5): 69-73.

相关系数对软件可靠性验证测试的影响分析

Analysis of Correlation Coefficient Influence on Software Reliability Demonstration Testing

  • 摘要: 针对在软件可靠性验证测试过程中传统贝叶斯方法存在着不区分历史信息与试验信息的问题, 提出了基于混合贝叶斯软件可靠性验证测试方案.引入相关系数概念, 分别通过主观定量法、拟合优度法、Kullback信息量法、Spearman相关系数法以及Kendall相关系数法确定混合贝塔分布中的相关系数.最后通过实验分析, 证明了这5种计算相关系数方法的可行性, 并且同时证明了区分历史信息和试验信息, 会对软件可靠性验证测试的可接收质量水平产生影响.

     

    Abstract: In the process of software reliability demonstration testing there is a problem that traditional Bayesian method did not distinguish between historical information and experimental information.A hybrid Bayesian software reliability demonstration testing scheme is proposed.By introducing the concept of correlation coefficient, the correlation coefficients in hybrid Beta distribution are determined by subjective quantitative method, goodness of fit method, Kullback information method, Spearman correlation coefficient method and Kendall correlation coefficient method respectively.Finally, through experimental analysis, prove the feasibility of the 5 kinds of calculation method of correlation coefficient, and prove that differentiating the historical information and experimental information impacts acceptable quality level of software reliability demonstration testing.

     

/

返回文章
返回