Γ分布的K检验

K-TEST OF Γ DISTRIBUTION

基于前人提出的统计理论 ,就工程中常用的Γ分布给出了一种新的拟合优度检验方法———K检验。与Kolmogorov和Smirnov提出的D检验相比 ,K检验在上下两个小概率尾区有更好的检验能力 ,这对长寿命高可靠性结构设计具有重要意义。利用蒙特卡洛数值模拟方法 ,得到了Γ分布在尺度参数未知、形状参数已知时K检验的临界值。

A new goodness of fit test, K test,is developed for Γ distributions.Theoretically,the test power of K test method is higher than that of D test method,especially in the two tail regions of probabilistic density curves,which is important for high reliability design and analysis Utilizing Monte Carlo simulating techniques,the critical values of K test for Γ distributions with unknown scale parameters are presented