摘要
从统计学的角度来讨论一致性测试的可靠性 .将输入空间E划分为不相交的等价子空间 ,然后建立一种新的数学模型 ,利用数理统计的知识得到了“零错误”时P的置信上界 .并且得到了在大样本性质下 ,此置信上界只与输入测试例个数和置信系数相关的结论 .
Conformance testing is considered from a statistical point of view in this paper. The input space is parlitioned into homogeneous disjoint subdomains, and a new mathematical model is constructed to find the confidence upper bound with the knowledge of mathematical statistics. It is concluded that this upper bound is merely correlated with the amounts of input samples and the confidence coefficient. Finally, an example is given to validate the related theory.
基金
国家自然科学基金重大研究计划项目 (90 1 0 4 0 1 0 )
国家自然科学基金 (6 0 2 4 1 0 0 4 )
国家"86 3"计划(2 0 0 1AA1 2 1 0 1 6 )
国家"973"计划 (2 0 0 3CB31 4 80 1 )
