Application of Non parametric Statistical Methods in Partial Data Analysis
Download as PDF
DOI: 10.25236/ssehr.2023.029
Corresponding Author
Yaowen Hu
Abstract
This study focuses on exploring the application of non parametric statistical methods in skewed data analysis, aiming to address the challenges faced in processing skewed data. This article demonstrates the practical application of non parametric statistical methods through two experiments. In Experiment 1, we used KDE (Kernel Density Estimation) to analyze the distribution characteristics of skewed data and emphasized the impact of bandwidth selection on the estimation results. The experimental results reveal how to use different bandwidth values for KDE and provide practical guidance for selecting the appropriate bandwidth. In Experiment 2, we applied the Wilcoxon rank sum test to detect outliers in skewed data. Through this test, we successfully identified outlier samples and demonstrated the robustness and reliability of this method in skewed data analysis. Overall, this study emphasizes the key applications of non parametric statistical methods in skewed data analysis, which are expected to help explain the distribution characteristics of skewed data, identify outliers, and promote further research and application in the field of skewed data analysis.
Keywords
Non parametric statistical methods; Partial data analysis; Skewed data