A Quadratic Regression Approach Using Mathematical Goal Programming | ||||
مجلة البحوث التجارية | ||||
Article 26, Volume 47, Issue 2, April 2025, Page 304-324 PDF (591.05 K) | ||||
Document Type: تجاریة کل ما یتعلق بالعلوم التجاریة | ||||
DOI: 10.21608/zcom.2025.376982.1440 | ||||
![]() | ||||
Authors | ||||
Doaa Hassan ![]() | ||||
1كلية التجارة _جامعة الازهر فرع البنات بالقاهرة | ||||
2قسم الإحصاء التطبيقي ، كلية الاقتصاد والعلوم السياسية ، جامعة القاهرة، مصر | ||||
3قسم الإحصاء كلية التجارة ، جامعة الأزهر فرع البنات بالقاهرة | ||||
4قسم الإحصاء كلية التجارة - جامعة الأزهر فرع البنات بالقاهرة | ||||
Abstract | ||||
Goal programming is a method commonly used to solve multi-criteria decision problems in different areas, such as manufacturing, production planning, marketing, supply chains, healthcare, management science, environment and energy. Goal programming approach minimizes the absolute value of the deviations to reduce the effect of outliers. This study aims to use a suggested quadratic regression approach using goal programming by adopting the least absolute deviation method (L_1-norm) and the least square method (L_2-norm) to estimate the parameters. A simulation study was applied to evaluate the proposed approach's performance by generating data for dependent variables from the lognormal and Cauchy distributions with three different parameters and samples with different sizes. The results of the various simulation studies that are carried out to compare the efficiency of the least absolute deviation method and the least square method estimators, for various error distributions, indicate that the least absolute method is more efficient than the least square method estimator | ||||
Keywords | ||||
Goal Programming; Quadratic Regression; Mean Square Error; Least Square Method; Least Absolute Method | ||||
Statistics Article View: 50 PDF Download: 36 |
||||