首页 | 本学科首页   官方微博 | 高级检索  
     检索      

指数分布异常数据检验的Bayes方法
引用本文:肖浩,刘次华,董锐,黄丽琨.指数分布异常数据检验的Bayes方法[J].华中科技大学学报(自然科学版),2004,32(8):92-94.
作者姓名:肖浩  刘次华  董锐  黄丽琨
作者单位:华中科技大学,数学系,湖北,武汉,430074
基金项目:国家自然科学基金资助项目 (70 2 71 0 6 9)
摘    要:研究了指数分布场合异常数据的检验问题,提出了无信息先验分布下的Bayes检验方法.利用分布参数的先验分布严格推导出新的检验统计量及其后验分布函数.实例和蒙特卡罗模拟表明了方法是可行的,适用于同时含有异常大、异常小数据的情形,并能够避免遭受屏蔽效应.

关 键 词:异常数据  指数分布  Bayes方法  先验分布  数据检验
文章编号:1671-4512(2004)08-0092-03
修稿时间:2003年11月27

Test of outliers by Bayesian
Xiao Hao Postgraduate, Dept. of Mathematics,Huazhong Univ. of Sci. & Tech.,Wuhan ,China. Liu Cihua Dong Rui Huang Likun.Test of outliers by Bayesian[J].JOURNAL OF HUAZHONG UNIVERSITY OF SCIENCE AND TECHNOLOGY.NATURE SCIENCE,2004,32(8):92-94.
Authors:Xiao Hao Postgraduate  Dept of Mathematics  Huazhong Univ of Sci & Tech  Wuhan  China Liu Cihua Dong Rui Huang Likun
Institution:Xiao Hao Postgraduate, Dept. of Mathematics,Huazhong Univ. of Sci. & Tech.,Wuhan 430074,China. Liu Cihua Dong Rui Huang Likun
Abstract:The paper studied the traditional tests of an exponential sample and presented a Bayesian test based on the noninformative prior distribution. By using prior distribution of the distributional parameter, test statistics and their post distributional functions were derived from the strict deduction. The example and Monte carlo simulation experiments showed its effectiveness. The test was suitable for the data including upper and lower outliers and could avoid the masking effect of the data.
Keywords:outliers  exponential distribution  Bayesian method  prior distribution
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号