A New Efficient Redescending M-Estimator for Robust Fitting of Linear Regression Models in the Presence of Outliers

Table 5

Total mean square error for contaminated datasets generated from normal distribution (80% X & Y∼N(0, 1) and 20% X∼N(20, 1) and 20% Y∼N (100, 1)).

N	P	OLS	Huber	Hampel	Andrew	Tukey	Ali
N = 30	P = 2	5.050	5.092	0.297	0.312	0.314	0.201
	P = 3	3.600	3.703	0.445	0.432	0.504	0.410
	P = 6	2.509	2.570	1.133	1.125	1.124	1.101
N = 50	P = 2	4.989	5.012	0.253	0.243	0.233	0.210
	P = 3	3.598	3.624	0.330	0.330	0.335	0.290
	P = 6	2.382	2.426	0.490	0.505	0.505	0.454
N = 100	P = 2	5.053	4.985	0.155	0.159	0.159	0.153
	P = 3	3.659	3.567	0.197	0.199	0.199	0.191
	P = 6	2.509	2.318	0.284	0.286	0.287	0.281