...
  • 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 robust Bayesian estimate of the concordance correlation coefficient
【24h】

A robust Bayesian estimate of the concordance correlation coefficient

机译：一致性相关系数的鲁棒贝叶斯估计

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

For development of drugs, evaluation of agreement of different methods for biomarker quantification is an important step and concordance correlation coefficient (CCC) is one of the most popular scaled indices for this purpose. Originally proposed by Lin (Ref. 1) two quantify the closeness of two paired observations such as methods, instruments, assays etc. to the 45° line that passes through the origin, called concordance line. This can be generalized for multiple observers also such as d observers with μ as the mean vector and ∑ as the covariance matrix. Let σ_i~2 be the variance and μ_i be the mean of the measurements made by the i~(th) observer and σ_(ij) be the covariance between measurements from observer i and j. If the CCC value is less than 0.65, then the agreement is considered as poor. The CCC value has to be at least 0.8 to have substantial agreement. There are many extensions available to CCC under various assumptions. This work was motivated by a real-world problem of analysis of EEG data that shows fairly good agreement but having low point a estimate and confidence intervals of CCC that, the point estimate was 0.7 and the lower bounds were below 0.5. This may be due to outliers or present in the data. Hence it was felt necessary to have a robust estimate of CCC that with stands outliers. This paper addresses the requirement. Available methods for this purpose lack precision and accuracy. If the distance functions are different then it is difficult to define cutoff points. Hence the applications were confined to only two observers. This paper proposes a CCC based on its original definition but that can overcome the shortcomings of the existing CCCs. (29 refs.)
机译：对于药物开发而言，评估生物标志物定量方法的一致性是一个重要步骤，一致性相关系数（CCC）是为此目的最受欢迎的定标指标之一。最初由Lin（参考文献1）提出，用于量化两个成对观测值（例如方法，仪器，测定法等）与穿过原点的45°线（称为一致线）的接近程度。这可以推广到多个观察者，例如d个观察者，其中μ为平均向量，∑为协方差矩阵。设σ_i〜2为方差，μ_i为第i〜（th）个观察者进行测量的平均值，σ_（ij）为来自观察者i和j的测量之间的协方差。如果CCC值小于0.65，则该协议被认为是差的。 CCC值必须至少为0.8才能达成实质性协议。在各种假设下，CCC都有许多扩展名。这项工作是由对EEG数据进行分析的一个现实世界问题激发的，该问题显示出相当好的一致性，但对CCC的估计值较低，并且CCC的置信区间为0.7，点估计值的下限低于0.5。这可能是由于异常值或数据中存在的值。因此，认为有必要对CCC进行可靠的估算，并与异常值进行比较。本文解决了这一要求。为此目的可用的方法缺乏准确性和准确性。如果距离函数不同，则很难定义截止点。因此，申请仅限于两名观察员。本文基于其原始定义提出了一种CCC，但它可以克服现有CCC的缺点。 （29参考）

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