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

Table 5

Prediction performance results over the testing phase.

	R^2	RMSE	MAE	MAPE	Nash	MD
MARS model
M1	0.75562	55636.4	39175.3	0.11317	0.75561	0.77964
M2	0.78542	52169	36042.4	0.10461	0.78513	0.79856
M3	0.81853	47949.3	36928.8	0.10698	0.81848	0.79537
M4	0.86669	41152.6	32861.1	0.09655	0.86629	0.82096
M5	0.94125	27422.7	19761.8	0.05454	0.94063	0.89374
M6	0.94125	27422.7	19761.8	0.05454	0.94063	0.89374
ELM model
M1	0.70389	61662	43434.8	0.12204	0.69981	0.75368
M2	0.72687	59271.2	41879.9	0.11948	0.72264	0.76358
M3	0.77521	53666.1	41776.6	0.11991	0.77262	0.76784
M4	0.80047	50554.1	40050.8	0.11642	0.79822	0.77885
M5	0.894	37239.8	25507.3	0.06922	0.89051	0.86168
M6	0.89463	37141.8	25376.6	0.06887	0.89109	0.86241
PLS model
M1	0.6997	62852.2	45250.3	0.12747	0.68811	0.74448
M2	0.71411	60628.3	43182.2	0.12358	0.70979	0.75411
M3	0.77381	53854.1	41631.9	0.11941	0.77102	0.76864
M4	0.80047	50554.3	40051.8	0.11643	0.79822	0.77884
M5	0.8941	37229.5	25541	0.06928	0.89057	0.86148
M6	0.89464	37140.3	25418.7	0.06898	0.89109	0.86217