Application of machine learning algorithms and feature selection in rapeseed (Brassica napus L.) breeding for seed yield

Table 7 Sensitivity analysis of the input features on the seed yield of rapeseed

Algorithm	Eliminated trait from inputs	R²	RMSE	MAE
MPLNN-Identity	–	0.838	0.270	0.214
	DPM	0.804	0.297	0.231
	PMB	0.833	0.275	0.218
	PP	0.836	0.272	0.215
	PAB	0.837	0.271	0.214
	BP	0.838	0.271	0.214
NuSVR-QP	–	0.871	0.241	0.195
	DPM	0.853	0.257	0.205
	SP	0.862	0.249	0.197
	FP	0.864	0.247	0.197
	PP	0.867	0.245	0.197
	TSW	0.867	0.245	0.197
MLR	–	0.846	0.263	0.208
	DPM	0.810	0.292	0.231
	PH	0.844	0.265	0.211
	BP	0.845	0.265	0.209
	PP	0.845	0.264	0.209
	DSF	0.846	0.263	0.208

R² determination coefficient, RMSE root mean square error, MAE mean absolute error, MLR multiple linear regression, NuSVR nu-support vector regression, MLPNN multilayer perceptron neural network, QP quadratic polynomial, PH plant height, PMB pods per main branch, PAB pods per axillary branches, PP pods per plant, BP branches per plant, DSF days to start of flowering, DPM days to physiological maturity, FP flowering period, TSW thousand seed weight, SP seeds per pod

ISSN: 1746-4811