多元函数型数据两样本成对检验问题研究

Two-Sample Paired Testing Problem for Multivariate Functional Data

  • 摘要: 本文针对多元函数型数据两样本成对检验问题展开研究,首先基于L2范数和F检验的思想,通过取积分和求上确界的方法提出了四种新的全局检验统计量,并推导其渐近零分布的形式,然后通过非参数自助法和Welch-Satterthwaite卡方近似法逼近检验统计量的渐近零分布,还推导了所提检验方法的√n/span>相合性.最后,通过数值模拟实验和空气污染数据对所提各种检验方法的有限样本性质进行了实证分析.

     

    Abstract: This paper addresses the problem of two-sample paired testing for multivariate functional data. Initially, four novel global test statistics are proposed based on the concepts of the L2 norm and the F-test, utilizing integration and supremum techniques, and their asymptotic null distributions are derived. The asymptotic null distributions of these test statistics are then approximated using the non-parametric bootstrap method and the Welch-Satterthwaite chi-square approximation method. Additionally, the √n-consistency of the proposed testing methods is established. Finally, an empirical analysis of the finite sample properties of the proposed test methods is conducted through numerical simulations and real data from air pollution.

     

/

返回文章
返回