广义Ⅱ型逐次截尾数据下逆高斯分布参数的极大似然估计

Maximum Likelihood Estimation of the Parameters of Inverse Gaussian Distribution Under General Progressive Type-Ⅱ Censoring

针对可靠性应用研究中常需要确定寿命分布参数的问题,讨论了在广义Ⅱ型逐次截尾样本下逆高斯分布位置参数和尺度参数的最大似然估计.通过参数变换得到了基于新参数的逆高斯分布模型,并利用牛顿迭代算法得到新参数的最大似然估计,进而求出原参数的极大似然估计.借助随机模拟方法和一个实例评价了最大似然估计.说明提出的极大似然估计理论具有应用价值.

It＇s well known that researchers often need to determine the life distribution parameters in the applications of reliability theory. In this paper, the maximum likelihood estimation （MLE） for the location and scale parameters of inverse Gaussian distribution under general progressive type-Ⅱ censoring was dicussed, An IG distribution model with new parameters was obtained by transformation and the MLEs of these new parameters were calculated using the Newton Raphson algorithm. Thereby it was possible to determine the MLEs of the original parameters. Furthermore, the proposed MLEs were evaluated by using the stochastic simulation data and real data. The obtained results showed that the maximum likelihood estimation theory suggested in the paper has an application value.