طرق إزالة أثر المشاهدات الشاذة على نموذج الانحدار الخطي البسيط

الملخص EN

In this sientific paper to eliminate the outliers impact on the simple linear regression models we proposed a new ways, wich inclowded a way for detecting outliers observation followed by a graphical measures so as to determine it’s figuers, and a way to treat the outliers values substitude them with new ones.

application has been done on 35 simple linear models targeting it’s compairement by the brevious ways and rational proficiency mean test of the proposed ways.

results showed that proposed ways are more perfect than the brevious ones as they increase the 2 r value and decrease MSE value, and lead to a significant development in the indicators of linear regression models.