Fig. 5

Box plots of measured and predicted SY of rapeseed using all measured traits as inputs. Algorithms are sorted based on the highest to lowest R² value from left to right. A. Training stage. B. Testing stage. MLR multiple linear regression, BRR Bayesian ridge regression, ARD automatic relevance determination, GLM generalized linear model, SGD stochastic gradient descent, NuSVR nu-support vector regression, ESVR epsilon support vector regression, LSVR linear support vector regression, MLPNN multilayer perceptron neural network, RBF radial basis function, QP quadratic polynomial, CP cubic polynomial, EI epsilon insensitive, SEI squared epsilon insensitive