Abstract
In this paper, we proposed new estimation methods called (RVIF) to address the common problem of multicollinearity and outliers. For this we propose to use strong coefficient determination (RR2) instead of the classical R2 to obtain the strong variance inflation vector ( RVIF). In order to evaluate the performance of the proposed method, we compared it with the existing methods. The results indicated that the proposed method has a high performance compared to other methods for all cases of sample size, linear relationship ratios, and contamination percentage
