首页> 外文会议>International Conference on Applied Statistics >An empirical likelihood ratio based goodness-of-fit test for some lifetime distributions
【24h】

An empirical likelihood ratio based goodness-of-fit test for some lifetime distributions

机译:基于经验似然比的一些寿命分布的健美测试

获取原文

摘要

In this paper, an empirical likelihood ratio based goodness-of-fit test for three important lifetime distributions is proposed. The three distributions are a generalized exponential distribution, a Weibull distribution, and a lognormal distribution. For each distribution, the proposed test statistic and three other test statistics are examined, namely the Kolmogorov-Smirnov test statistic, the Cramer-von Mises test statistic and the Anderson-Darling test statistic. Method for parameter estimation in this study is a maximum likelihood. Some investigations on type I error control and power of the test of these statistics are studied by simulation. The selected sample sizes are 10, 25, 50 and 100. The simulation studies show that the proposed test statistics with the optimal values of δ which locate below or around 0.5 can control type I error well. Power studies using several different distributional forms show that the proposed test statistics are competitive when compared with other available test statistics.
机译:在本文中,提出了一种基于基于拟合寿命分布的拟合优度测试的经验似然比。三个分布是广义指数分布,威布尔分布和逻辑正式分布。对于每个分布,审查所提出的测试统计和三个其他测试统计数据,即Kolmogorov-Smirnov测试统计数据,Cramer-von Mises测试统计和Anderson-Darling测试统计数据。本研究中参数估计的方法是最大可能性。通过模拟研究了关于I型错误控制和测试的电源的一些调查。所选的样本大小为10,25,50和100.模拟研究表明,所提出的测试统计数据与δ定位的Δ的最佳值可以控制I误差。使用几种不同的分布形式的功率研究表明,与其他可用测试统计数据相比,所提出的测试统计数据具有竞争力。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号