赵丽华, 刘桂芬, 刘志永, et al. 基于共轭先验分布的两小样本Bayesian检验及实例分析[J]. 2011, (1).DOI:
基于共轭先验分布的两小样本Bayesian检验及实例分析
摘要
介绍基于共轭先验分布的两小样本Bayesian检验原理及方法
并与经典方法作对比研究.当两组资料确实存在差别且为小样本时
基于共轭先验分布的两小样本Bayesian检验方法比传统的两样本t检验法得到的置信区间范围要短
精度要高.本文通过实例分析验证了此结论.
Abstract
The theory and method of Bayesian test for two small samples based on conjugate prior distribution were introduced and compared with the classical method.While there exists surely difference between two parameters which are of small samples
conclusions drawn from Bayesian are more steadily than that from classical t test and can be interpreted reasonably.By comparing the Bayesian test with the classical t-test for two samples
the confidence interval of the former is shorter
and the accuracy is higher.The conclusion was verified by the instance analysis.