2024, Vol. 5, Issue 1, Part B
Developing new capabilities for the community mean in the presence of extreme values: A comparative study
Author(s): Omar Makki Rahim and Sazan Najat Najmuddin Sadiq
Abstract: This study aims to develop a dynamic simulation framework to evaluate the performance of several statistical estimators under different data conditions, including clean data, contaminated data with extreme outliers, and heavy-tailed distributions such as t(3)t(3)t(3). Eight estimators (Mean, Median, Trim, Winsor, Huber, Tukey, HL, and the proposed TA-Mean) were assessed through a comprehensive simulation experiment using a sample size of n=50n=50n=50 and contamination levels ranging from 0% to 20%. The results showed that the arithmetic mean deteriorates sharply under contamination, while robust estimators maintained stable performance. The proposed TA-Mean achieved the best balance between robustness and efficiency across all scenarios. These findings support the four hypotheses regarding the impact of contamination and distributional shape on estimator accuracy and recommend using TA-Mean in environments with contaminated or heavy-tailed data.
DOI: https://doi.org/10.22271/math.2024.v5.i1b.248
Pages: 191-199 | Views: 502 | Downloads: 143
Download Full Article: Click Here
How to cite this article:
Omar Makki Rahim and Sazan Najat Najmuddin Sadiq. Developing new capabilities for the community mean in the presence of extreme values: A comparative study. Journal of Mathematical Problems, Equations and Statistics. 2024; 5(1): 191-199. DOI: 10.22271/math.2024.v5.i1b.248
