• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文学位 >Joint modeling of mean and variance using a generalized quasi-likelihood function
【24h】

Joint modeling of mean and variance using a generalized quasi-likelihood function

机译：使用广义拟似然函数对均值和方差进行联合建模

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

The mean and variance are jointly modeled using a generalized quasi-likelihood function (GQL). This function does not make any underlying assumptions of a distribution to model the mean or the variance. In using this function there are two nonlinear parameters that need to be estimated. These parameters are essential in determining the nature of the variance function. Consistent and simple estimators are derived for these parameters.;The modeling of the mean and the variance consists of an extended quasi-likelihood function that assumes a gamma distribution for the modeling of the variance with a square variance function. This method involves one nonlinear parameter for shaping the nature of the variance function only in the mean submodel and makes use of the profile quasi-likelihood method to estimate the parameter.;The presentation of the generalized quasi-likelihood in this dissertation relaxes the restrictions placed on the variance function in the variance of the mean submodel. It makes no underlying distributional assumption and presents an alternative to the profile quasi-likelihood for nonlinear parameter estimation. The nonlinear parameters in this generalization are obtained through weighted least squares and the method of moments. This method of estimation provides consistent and efficient estimators. Asymptotic properties of these estimators are obtained.;The Ames Salmonella Reverse Mutagenicity Assay data are analyzed and the results are compared with that of other generalizations and extended quasi-likelihood which use profile quasi-likelihood method of estimation for nonlinear parameter estimation. The GQL function is used to analyze data on absence from school from a sociological study of Australian Aboriginal and white children and data obtained to monitor the performance of its accredited asbestos fibre counters.
机译：均值和方差使用广义拟似然函数（GQL）联合建模。该函数不对分布进行任何基础假设来对均值或方差建模。使用此功能时，需要估计两个非线性参数。这些参数对于确定方差函数的性质至关重要。为这些参数导出一致且简单的估计量。均值和方差的建模由扩展的拟似然函数组成，该函数假设伽玛分布用于具有方差函数的方差建模。该方法只涉及一个非线性参数，用于仅在均值子模型中塑造方差函数的性质，并利用轮廓拟似然法对参数进行估计；本文对广义拟似然的表述简化了约束。关于均值子模型方差的方差函数。它没有基础的分布假设，并且为非线性参数估计提供了轮廓拟似性的替代方法。这种概括中的非线性参数是通过加权最小二乘和矩量法获得的。这种估算方法可提供一致且有效的估算器。分析了Ames沙门氏菌反向诱变分析的数据，并将结果与​​其他概化和扩展的拟似然性进行了比较，后者使用轮廓拟似然估计法进行非线性参数估计。 GQL函数用于分析澳大利亚原住民和白人儿童的社会学研究中失学的数据，以及监测其认可的石棉纤维计数器性能的数据。

著录项

  • 作者

    Khanna, Sarita.;

  • 作者单位

    Arizona State University.;

  • 授予单位 Arizona State University.;
  • 学科 Statistics.;Biostatistics.;Industrial engineering.
  • 学位 Ph.D.
  • 年度 1996
  • 页码 108 p.
  • 总页数 108
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

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