استخدام مؤشر متوسط مطلق البواقي المعيارية كمؤشر مقترح للكشف عن جودة نموذج الانحدار الخطي

Abstract EN

The study targeted detecting quality linear regression model using the proposed indicator, an indicator of average absolute residuals standard.

To achieve this goal a random sample of students data from faculty of education, King Saud University in ten courses of the joint between the science departments college as independent variables (predictors) and variable grade point average (GPA) as the dependent variable predictable.

The study also used a set of standards that are used in the comparison between the models of linear regression like a Akiak information criterion (AIC) and Bayesian information criterion (-BIC) and the adjusted coefficient of determination debugger (R2), and standard error (S), statistic Mallows (Cp), standard (PRESS) as criteria indicators to test the proposed index can be used as an alternative to the procedures in the trade off between regression models resulting from the multiple linear regression analyses depending on the method (Stepwise) input variables in the regression model.

The results showed the possibility of using the suggested index in the trade-off between regression models.