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

Table 3

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	50	100	100	50	0.0	0.0	0.3	32.3	0.0
	40	45	95.6	100	55	4.4	0.0	0.0	19.2	0.0
	100	20	95.2	100	80	4.4	0.0	1.0	18.7	0.0
	200	20	96	100	80	4.0	0.0	2.4	13.6	0.0
10%	20	0.0	90	100	100	10	0.0	0.0	26.7	1.5
	40	25	82	99	75	18	1.0	0.0	17.5	1.0
	100	30	84.4	97.4	70	13.6	2.6	1.0	17.0	1.0
	200	30	88	98	70	12	2.0	2.0	9.4	1.5
15%	20	0.0	77.2	100	100	22.8	0.0	0.0	26.7	0.0
	40	14	84.2	94	86	15.8	6.0	0.0	18.6	2.5
	100	14	79	94.2	86	21	5.8	0.0	15.0	2.0
	200	7	82.1	98	93	17.9	2.0	1.5	6.5	2.0
20%	20	0.0	75.3	99	100	24.7	1.0	0.0	16.7	0.0
	40	12	67.9	93.3	88	32.1	6.7	0.0	11	2.5
	100	10	63	93.1	90	37	6.9	0.0	10.0	2.0
	200	8	70	97.1	92	30	2.9	1.0	8.0	1.5