Abstract

Computational data analysis is an essential component of modern statistics. The most important challenge in classical statistics was an estimation of the location parameter. In addition, the selection of estimators among potential estimators is another problem in the computational data analysis. Therefore, the main objective of this study is to develop a location estimator that has been modified compared to all common location estimators. In order to ensure the objective of the new estimator, combinations weighted mean (Alex-mean) was developed and computed with R version 3.6.3. Further, the properties of a good location estimator diagnosed with all location estimators compared to Alex-mean with a COVID-19 new case data set. In addition, the new estimator was compared with mean, median and trimmed-mean based on relative efficiency and bootstrap simulation techniques. The new estimator is better than the mean, median and trimmed-mean of relative efficiency estimated using bootstrap and simulation techniques. The sensitivity graph shows that Alex-mean is insensitive to the outlier and is computed on the basis of all observations. Thus, the researcher suggested that Alex-mean is the good location estimator than all common measures of central tendency since it is undisturbed by extreme values.

摘要

计算数据分析是现代统计学的重要组成部分。经典统计学中最重要的挑战是位置参数的估计。此外，潜在估计量中估计量的选择是计算数据分析中的另一个问题。因此，本研究的主要目标是开发一种与所有常见位置估计器相比经过修改的位置估计器。为了确保新估计器的目标，使用 R 版本 3.6.3 开发并计算了组合加权平均值 (Alex-mean)。此外，与具有 COVID-19 新病例数据集的 Alex-mean 相比，使用所有位置估计器诊断的良好位置估计器的属性。此外，新的估计量与平均值进行了比较，基于相对效率和引导模拟技术的中值和修整平均值。新的估计器优于使用引导程序和模拟技术估计的相对效率的均值、中值和修整均值。敏感性图显示 Alex 均值对异常值不敏感，并且是根据所有观测值计算的。因此，研究人员认为 Alex-mean 是比所有常见的集中趋势度量更好的位置估计器，因为它不受极值的干扰。