مقارنة طريقة Lasso مع طريقة Lasso Multi-split لتقدير انموذج الانحدار متعدد المتغيّرات في ظل بيانات عالية الأبعاد باستخدام المحاكاة
Year: 2024
Type: article
Abstract: This study aims to compare some of the organizing methods used in estimating a multivariable regression model in the presence of high-dimensional data. A comprehensive analysis of the data was conducted, studying the relationship between the independent variables and the dependent variable. These methods were applied to a set of real data to evaluate their performance and effectiveness. A comparison was made between the Lasso method, which is used for high-dimensional data, to find estimators in the case of dimensionality problem, and the Multi-split Lasso method. The estimation results were analyzed, and the accuracy and overall performance of each method were evaluated. The study reached important results indicating that the Multi-split Lasso method is superior to the Lasso method, based on several criteria, most notably the BIC criterion, which favored the proposed method for the regression model in high-dimensional data.
Authors: Luma Kais, Tariq Aziz
Language: English
Subfield: Statistics and Probability
Field: Mathematics
Domain: Physical Sciences
Open Access status: diamond
OpenAlex ID: W4400602729