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

Table 5 The performance of machine learning algorithms using selected traits by feature selection methods as inputs

Algorithm	Feature selection method	Training			Testing
Algorithm	Feature selection method	R²	RMSE	MAE	R²	RMSE	MAE
Multiple Linear Regression (MLR)	Lasso/FS	0.833	0.266	0.207	0.828	0.295	0.225
Ridge Regression (RR)	Lasso/FS	0.829	0.269	0.208	0.837	0.288	0.224
Generalized Linear Model (GLM)	SS/BS	0.834	0.265	0.212	0.842	0.283	0.225
Nu-Support Vector Regression (NuSVR)/Radial Basis Function (RBF)	SS/BS	0.845	0.256	0.200	0.830	0.293	0.228
	Lasso/FS	0.833	0.266	0.201	0.837	0.288	0.219
Epsilon Support Vector Regression (ESVR)/Linear	Lasso/FS	0.828	0.269	0.209	0.839	0.286	0.224
Epsilon Support Vector Regression (ESVR)/Sigmoid	SS/BS	0.504	0.459	0.347	0.541	0.483	0.376
Epsilon Support Vector Regression (ESVR)/Cubic Polynomial	SS/BS	0.245	0.566	0.430	0.311	0.592	0.488
	Lasso/FS	0.417	0.497	0.380	0.570	0.468	0.387
Multilayer Perceptron Neural Network (MLPNN)/Identity	SS/BS	0.827	0.270	0.219	0.843	0.283	0.224
	Lasso/FS	0.826	0.272	0.210	0.838	0.286	0.224
Multilayer Perceptron Neural Network (MLPNN)/Tanh	SS/BS	0.834	0.265	0.211	0.842	0.283	0.229
	Lasso/FS	0.828	0.269	0.208	0.839	0.286	0.224
Multilayer Perceptron Neural Network (MLPNN)/Relu	SS/BS	0.839	0.261	0.209	0.833	0.291	0.231

FS forward selection, SS stepwise selection, BS backward selection, R² determination coefficient, RMSE root mean square error, MAE mean absolute error

ISSN: 1746-4811