A New Robust Diagnostic Plot for Classifying Good and Bad High Leverage Points in a Multiple Linear Regression Model

Table 2

Percentage of correctly identified BLPs, masking, and swamping for simulation data ().

Cont. level		% Correct detection			% Masking			% Swamping
		OLS-MD	LMS-RMD	MGt-DRGP	OLS-MD	LMS-RMD	MGt-DRGP	OLS-MD	LMS-RMD	MGt-DRGP
5%	20	75	100	100	25	0.0	0.0	5.0	15.5	0.0
	40	50	100	100	60	0.0	0.0	2.2	14.8	0.0
	100	56	100	100	44	0.0	0.0	0.0	6.40	0.0
	200	40	100	100	60	0.0	0.0	0.9	4.60	0.0
10%	20	50	100	100	50	0.0	0.0	5.0	10.0	0.0
	40	50	90	100	60	10	0.0	0.0	17.5	0.0
	100	30	87.3	100	70	12.7	0.0	0.0	5.0	1.2
	200	30	92.3	100	70	7.7	0.0	1.0	4.0	1.0
15%	20	0.0	100	100	100	0.0	0.0	5.0	10.0	0.0
	40	16.7	66.7	97.0	83.3	33.3	3.0	0.0	15.0	0.0
	100	13.3	70	96.2	86.7	30.0	3.8	0.0	7.30	1.1
	200	23.3	85	99	76.7	15.0	1.0	0.0	2.0	1.5
20%	20	0.0	50	100	80	50	0.0	5.0	23.0	0.0
	40	12.5	62.5	95.7	81.5	38.5	4.3	0.0	15.0	2.1
	100	15	67	97.1	85	33.0	2.9	0.0	8.20	0.8
	200	20	82	98.1	80	18	1.9	1.0	3.70	0.5