微波效应实验最小样本量估计分析  

Estimating analysis of least sample of microwave effects experiment

作　　者：李科[1] 马弘舸[1] 周海京[2] 

机构地区：[1]中国工程物理研究院应用电子学研究所,四川绵阳621900 [2]北京应用物理和计算数学研究所,北京100088

出　　处：《强激光与粒子束》2009年第4期541-544,共4页High Power Laser and Particle Beams

基　　金：国家高技术发展计划项目

摘　　要：为规范微波效应实验最小样本量,采用区间估计的统计推断方法,开展了样本作用概率与总体作用概率对应关系的分析和研究工作,对满足一定置信度和一定总体作用概率下所需要的最小样本量进行估计。研究结果表明,实验最小样本量确定为10是较为合理的。在80%的置信度下,当10支样本量样本作用概率达到70%时,总体作用概率置信下限约为51.6%。通过该方法能够将所获取的微波效应数据通过同一指标"总体作用概率置信下限"进行有效整理。Seeking for the criterion of the least sample of microwave effects experiment, the research work between sample effect probability and collectivity effect probability is developed by the interval estimating. The Bernoulli distribution of a single microwave effect data is convered to the beta distribution, then expressed by the fractile of the F distribution, so the relational ex- pression relating experimental sample, eonfidenee level, sample effect probability and the collectivity effect probability is derived. The computing result indicates that it is comparatively reasonable when the sample of microwave effects experiment is 10, thus, the confidence lower limit of collectivity effect probability is about 51.6% in 80 % confidence level when the sample effect probability is 70%. In this way, the previous data of microwave effects can be rearranged by means of the confidence lower limit of collectivity effect probability.

关 键 词：微波效应 最小样本量 概率 估计 置信下限 

分 类 号：TN015[电子电信—物理电子学]