Evaluation of Several Machine Learning Models for Field Canal Improvement Project Cost Prediction

Table 4

Prediction performance results over the training phase.

	R^2	RMSE	MAE	MAPE	Nash	MD
MARS model
M1	0.76	57097.21	40090.04	0.11	0.76	0.78
M2	0.79	52443.11	35922.94	0.10	0.79	0.80
M3	0.83	46957.45	35377.61	0.10	0.83	0.81
M4	0.87	41427.33	31383.75	0.09	0.87	0.83
M5	0.94	28458.17	20444.55	0.06	0.94	0.89
M6	0.94	28458.17	20444.55	0.06	0.94	0.89
ELM model
M1	0.72	61144.00	42644.28	0.12	0.72	0.76
M2	0.75	58309.91	40861.50	0.12	0.75	0.77
M3	0.79	52776.65	40372.48	0.11	0.79	0.78
M4	0.81	50951.98	39296.32	0.11	0.81	0.79
M5	0.90	36019.94	24352.69	0.07	0.90	0.87
M6	0.90	36011.43	24362.29	0.07	0.90	0.87
PLS model
M1	0.72	61182.31	43650.76	0.12	0.72	0.75
M2	0.73	59634.10	42063.22	0.12	0.73	0.76
M3	0.79	52816.73	40059.10	0.11	0.79	0.78
M4	0.81	50951.98	39297.60	0.11	0.81	0.79
M5	0.90	36026.97	24406.13	0.07	0.90	0.87
M6	0.90	36013.16	24393.67	0.07	0.90	0.87