Determination the effects of outliers at the least squares, Ridge regression and Robust regression analysis results

Least square method given the best prediction if the assumption of multiple linear regression analysis was provided. Howewer there is any linear relation between independent variables and outliers in the data group, the assumptions can?t be provided. In this case regression coefficients ?t? values, ?F? values, error of mean square and coeeffcient of determination are changed.Multicollinearity and outliers effect are investigated in this study. Ridge regression (RR) method is suggested for multicollinearity and some outlier calculation method for outliers. Parameter prediction are realized both in outliers existance and nonexistance via Least square, Ridge and M- estimatior. R2 value, error of mean square, ?F? nad ?t? values are used as comparision criteria.