岩土工程参数Bayes法推断中先验与后验样本大小的探讨  

Discussion of Priori and Posteriori Sample's Number of the Geotechnical Parameters with the Bayes Inference Method

作　　者：曾晟[1] 徐华清[1] 张佳男 马青叶 

机构地区：[1]南华大学核资源工程学院,湖南衡阳421001 [2]广东悉筑建筑设计有限公司,广东茂名525000 [3]云南金沙矿业股份有限公司汤丹公司,云南昆明650000

出　　处：《南华大学学报（自然科学版）》2015年第3期30-33,共4页Journal of University of South China：Science and Technology

摘　　要：岩土力学参数Bayes统计推断存在验前和后验样本信息问题．在解决工程问题时，先验与后验分布信息的获取是靠样本容量的大小确定，而获取样本信息是要付出工程费用的．为了探讨岩土参数Bayes法推断中先验与后验样本的取值大小，采用贝叶斯推断简化方法，通过对某工程的岩土力学参数分析，得到先验样本数为30～35个和后验样本数不少于4个时，计算出的统计参数均值与方差的误差最小．本文的研究为岩土工程力学参数概率统计样本大小的确定提供了参考．There are some information problems between priori and posteriori sample＇s number according to the Bayes inference method for geotechnical parameters. When solving the engineering problems, the acquisition of priori and posteriori distribution is determined by the number of the sample, and the project fee must be paid while getting the sample number. The simplified Bayesian inference method has been put to apply to analyzing the priori and posteriori sample number of geotechnical parameters. According to the analysis of geotechnical parameters, the conclusion has been gotten as following. The calculated error of statistical parameters＇ mean and variance is minimum when the number of priori sam- ples and posteriori samples reaches to 30 -50 and posteriori sample number is no less than 4. This dissertation provides references for the determination of geotechnical engineering mechanics parameters of sample number of probability and statistics.

关 键 词：岩土力学参数 Bayes法 先验样本 后验样本 样本大小 

分 类 号：TU43[建筑科学—岩土工程]