...
  • 主题
高级搜索  >
历史搜索 清空历史
首页> 外文期刊>Quality Control and Applied Statistics >A diagnostic for assessing the influence of cases on the prediction of missing data
【24h】

A diagnostic for assessing the influence of cases on the prediction of missing data

机译：用于评估案例对丢失数据预测的影响的诊断程序

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 following introductory remarks are explanatory of the background context: ·A critical aspect of statistical modeling is the identification of cases that have a significant influence on certain inferential results. ·Such cases may: —Indicate recording errors or anomalies in the phenomenon that produced the data. —Serve as an indication that the underlying model is too simplistic. ·Hence,the problems of model selection and influence case deletion must be adressed simultaneously. ·Examples of popular influence diagnostics (or sometimes referred to as deletion diagnostics) include Cook's distance, DFBETAS,DFFITS and COVRATIO. ·The purpose of these diagnostics is to project or highlight the effect of a case on a specific inferential objective. ·Typical instances of inferential quantities are regression parameter estimates,fitted values and estimated generalized variances. ·The deletion diagnostics compare inferential quantities based on fitting a model to the full data set wtih those based on fitting a model to the data set with a specific case removed. ·Two objectives are commonly observed in the context of statistical modeling: —To estimate the model parameters in the presence of missing data. —To impute reasonable values for the missing observatins. ·In the case later objective,the paper proposes a diagnostic that: —Assesses the influence of a case on the prediction of missing entries. —Is often conveniently evaluated in the setting of the EM algorithm. The utility and effectiveness of the proposed method was illustrated using examples of two real data sets: ·DS1:Bivariate data of RBI (runs batted in) versus home runs for the 105 National League players (1998 season). ·DS2:A time series consisting of cardiovascular mortality readings from the Los Angeles area.(16 refs.)
机译：以下介绍性说明可解释背景情况：•统计建模的关键方面是识别对某些推论结果有重大影响的案例。 ·这种情况可能是：—指示记录错误或产生数据现象的异常。 —表示基础模型过于简单。 ·因此，模型选择和影响案例删除的问题必须同时解决。 ·流行的影响诊断（或有时称为删除诊断）的例子包括库克距离，DFBETAS，DFFITS和COVRATIO。 ·这些诊断的目的是预测或强调案例对特定推论目标的影响。 ·推断量的典型实例是回归参数估计，拟合值和估计的广义方差。 ·删除诊断程序将基于将模型拟合到完整数据集的推断量与基于将模型拟合到数据集的特定案例被删除的推断量进行比较。 ·在统计建模的上下文中通常会观察到两个目标：—在缺少数据的情况下估计模型参数。 -为缺失的观察素估算合理的值。 ·针对案例的后续目标，本文提出了一种诊断方法：—评估案例对缺失条目的预测的影响。 -通常可以在EM算法的设置中方便地进行评估。使用两个真实数据集的示例说明了该方法的实用性和有效性：·DS1：RBI（打入的跑步）与本垒打的105个国家联赛球员（1998赛季）的双变量数据。 ·DS2：一个包含来自洛杉矶地区的心血管死亡率读数的时间序列。（16个参考）

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