利用样本空间排序法推断Logistic模型下的响应率

Inference for Response Probability Based on the Method of Ordered Relations in the Sample Spaces Under Logistic Model

  • 摘要: 基于 L ogistic响应模型 ,在感度数据下 ,应用样本空间排序法给出了响应率的下限估计 .在进行火炸药产品的响应率的测定时 ,利用感度实验 ,得到感度数据 (ni,si) ,i=1,2 ,… ,k的形式 ,将这种数据形式转换成样本空间中的序的形式 ,进一步利用样本空间中的序的理论给出响应率的下限估计 .在此基础上进行蒙特卡罗模拟 ,将其结果与渐近正态的方法进行比较 .并且将该方法用于 QD- 8电雷管的真实实验数据 .模拟和实例结果表明 ,在样本量较小时 ,应用样本空间排序法可以较好地估计响应率下限

     

    Abstract: Gives the lower limit estimation of response probability based on the method of ordered relations in sample spaces under logistic response model. To estimate the response probability of detonator products, a sensitivity test is always used. The sensitivity data (n i, s i), i=1,2,…,k can be obtained and will be transformed into the ordered relations in sample spaces. Using the theory of ordered relations in sample spaces, the lower limit estimation of response probability is given. Based on this study, the Monte Carlo simulation is given, compared with the result from the method of asymptotic normality, and is applied to the detonator QD 8. Numerical results show that the method of ordered relations in sample spaces can much better estimate the lower limit estimation of response probability under cases of small samples.

     

/

返回文章
返回
Baidu
map