Machine learning for high-throughput field phenotyping and image processing provides insight into the association of above and below-ground traits in cassava (Manihot esculenta Crantz)

Table 4 Root yield ML Model comparison

	ML Method	PCA			PCR			PCA vs PCR Difference (%)
	ML Method	R²	RMSE	RRMSE	R²	RMSE	RRMSE	R²	RMSE	RRMSE
Validation	RF	0.93	443.07	9.88%	0.94	449.09	9.19%	1.07%	1.35%	7.24%
	SVM	0.67	888.23	20.91%	0.63	947.45	22.10%	6.15%	6.45%	5.53%
	kNN	0.64	930.57	21.94%	0.64	953.3	22.01%	0.00%	2.41%	0.32%
	ANN	0.69	893.16	20.26%	0.7	910.21	21.12%	1.44%	1.89%	4.16%
Test	RF	0.66	889.28	20.52%	0.64	891.86	21.12%	3.08%	0.29%	2.88%
	SVM	0.64	916.77	21.23%	0.64	874.13	21.14%	0.00%	4.76%	0.42%
	kNN	0.67	899.8	20.10%	0.67	879.09	20.20%	0.00%	2.33%	0.50%
	ANN	0.61	1021.81	22.73%	0.61	1120.04	22.61%	0.00%	9.17%	0.53%

ISSN: 1746-4811