Table 1.

Corn yield prediction performance for years 2016, 2017, and 2018.

Model	Validation Year	Training RMSE	Training Correlation Coefficient (%)	Validation RMSE	Validation Correlation Coefficient (%)
CNN-RNN	2016	13.26	93.02	16.48	85.82
	2017	12.75	93.68	15.74	88.24
	2018	11.48	94.99	17.64	87.82
RF	2016	13.38	92.74	25.48	69.52
	2017	14.31	92.39	29.40	69.03
	2018	14.40	92.39	26.02	70.55
DFNN	2016	12.34	94.43	27.23	81.91
	2017	11.21	95.09	23.88	79.57
	2018	11.54	95.25	21.37	79.85
LASSO	2016	19.88	81.81	32.58	61.90
	2017	20.62	81.83	27.06	61.18
	2018	20.81	83.63	31.30	55.95

RF and DFNN stand for random forest and deep fully connected neural network, respectively. The average ± standard deviation for corn yield in years 2016, 2017, and 2018 are, respectively, 165.72 ± 30.35, 168.50 ± 32.88, and 170.77± 34.95. The unit of RMSE is bushels per acre.