Direct Interval Prediction of Landslide Displacements Using Least Squares Support Vector Machines

Table 3

CWC comparisons between the proposed and compared methods.

Cases			PSO-LSSVM	GA-LSSVM	Bootstrap-ELM-ANN	PSO-ELM
Baishuihe landslide	XD01	Best	0.96	1.10	1.18	1.90
			(19.79%)	(30.00%)	(34.75%)	(59.47%)
		Median	1.04	1.35	1.67	2.75
			(14.42%)	(34.07%)	(46.71%)	(67.64%)
		SD	0.67	0.21	0.12	0.37
			(89.55%)	(66.67%)	(41.67%)	(81.08%)
	ZG118	Best	0.94	1.08	1.50	2.38
			(12.77%)	(24.07%)	(45.33%)	(65.55%)
		Median	1.34	1.26	1.64	3.19
			(20.90%)	(15.87%)	(35.37%)	(66.77%)
		SD	0.44	0.60	0.15	0.98
			(65.91%)	(75.00%)	(0.00%)	(84.69%)
Tanjiahe landslide	ZG289	Best	0.14	0.21	0.34	13.25
			(21.43%)	(47.62%)	(67.65%)	(99.17%)
		Median	0.15	0.40	0.35	9015.04
			(13.33%)	(67.50%)	(62.86%)	(100.00%)
		SD	0.49	1.04	0.11	11166.22
			(87.76%)	(94.23%)	(45.45%)	(100.00%)
	ZG287	Best	0.12	0.11	0.24	1.22
			(25.00%)	(18.18%)	(62.50%)	(92.62%)
		Median	0.22	0.18	0.25	126.23
			(45.45%)	(33.33%)	(52.00%)	(99.90%)
		SD	0.06	0.94	0.08	35571.46
			(16.67%)	(94.68%)	(37.50%)	(100.00%)