Bootstrap方法的仿真实现及其在系统偏差估计中的应用

Simulation and Appication of Bootstrap Methods in System Error Estimation

该文研究Bootstrap方法与Bayes Bootstrap方法的原理及在小样本情况下的应用问题。分析了这二种统计方法的原理,给出了总体分布中未知参数的点估计方法和区间估计方法,后采用非参数抽样法,编写C语言程序,实现了Bootstrap方法与Bayes Bootstrap方法的计算机仿真。通过实例分析,给出了射击系统偏差在Bayes Bootstrap方法、Bootstrap方法和经典统计方法下的具体偏差值,得到Bayes Bootstrap方法优于其它二种方法的结论。

The theory of Bootstrap method and the Bayes Bootstrap method and the application of the two methods in the small sample are studied. The paper analyses the principle of the two methods and gives two estimation methods, namely, the point estimation and the interval estimation to reckon unknown parameters in population distribution. It implements the simulation of Bootstrap method and Bayes Bootstrap method by means of non-parameter sample and by compiling C language. By analyzing an example, it presents the concrete errors of the shot system warp under the Bayse Bootstrap method, the Bootstrap method and the classic statistics method. The conclusion is that the Bayes Bootstrap method is better than the other two methods.