摘要
几何分布是离散型寿命分布中最为重要的分布之一,许多产品的寿命(比如开关等)都可以用几何分布来描述。由于几何分布的无记忆性,它在可靠性理论与应用概率模型中有着非常重要的地位。目前,对关于几何分布在全样本场合、截尾样本场合以及加速寿命试验场合下参数的统计分析已经有了广泛的研究,并且有着重要的理论与应用价值。因此将不完全数据场合下的几何分布问题转化为指数分布问题,再利用指数分布的已有结果首次得到了几何分布在缺失数据场合和分组数据场合下参数的近似点估计,Monte-Carlo模拟算例结果令人满意,说明该方法是可行的。
Geometric distribution is one of the most important discrete life distributions. The hves of many products such as the switch can be described as the geometric distribution. Besides, the geometric distribution has played an important role in the reliability theory and the applied probability model because of its nonmemory property. The geometric distribution has widely studied on the statistical analysis of its parameter in the case of the full sample, the censored sample and the accelerated life test, which has obtained the important theory and application value. In this paper, the problem of geometric distribution with the incomplete data is translated into that of the exponential distribution. Then based on the missing data and grouped data with the geometric distribution, the approximate point estimates of parameters are firstly obtained by using the conclusions of exponential distribution. In addition, the results of some examples by Monte- Carlo simulations are satisfactory, which also illuminates the feasibility of the methods.
出处
《统计与信息论坛》
CSSCI
2010年第2期16-19,共4页
Journal of Statistics and Information
基金
上海师范大学科研项目<混合模型的可靠性统计分析>(SK200811)
上海市重点学科(S30405)
科学计算上海高等学校重点实验室
2009年上海市教委<概率论与数理统计重点课程>
上海师范大学第五期重点学科(DZL805)资助
关键词
几何分布
缺失数据
分组数据
点估计
geometric distribution
missing data
grouped data
point estimate
