基于数据融合的全生命周期软件可信性定量评估方法  

Quantitative Evaluation Across Software Development Life Cycle Based on Data Fusion

作　　者：张卫祥[1] 刘文红[1] 吴欣[1] 

机构地区：[1]北京跟踪与通信技术研究所,北京100094

出　　处：《计算机科学》2013年第11A期192-195,213,共5页Computer Science

摘　　要：提出了一种全生命周期的软件可信性定量评估方法。首先,建立全周期软件可信评估模型,对生命周期各阶段软件可信性进行逐层分解,分别设计定量或定性度量指标;然后,利用知识发现原理,获取软件可信特征树的权值分布;利用数据融合理论,对获得的多类型多量纲的可信度量数据进行分布式综合处理与推理;最后,给出软件可信性评估实例。工程实践表明,该方法能够有效保障软件评估过程的客观性和评估结果的准确度。This paper brings out a method on quantitative software trustworthy evaluation across software development life cycle. First, builds a hierarchical assessment model with decomposition of software trustworthy on the various stages of software development life cycle, designs appropriate quantitative or qualitative metrics set; then, using of knowledge discovery in database techniques to obtain the weights of all software trustworthy characteristics; finally, using of data fusion theory to process and reason large volumes of multi--type measurement data. Engineering practice shows that it can effectively improve objectivity of the assessment process and the accuracy of the assessment results.

关 键 词：软件定量评估 生命周期 数据融合 可信性 软件工程 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]