Fig 8. Complex interactions through nonlinear support vector machine (using the Gaussian kernel) boundary for the whole data with the leading eight influential variables, as identified in classification and regression tree relative variable importance plot (cf. Fig 6).

Four different categories: 1,2,3,4 (individual class) represent 1^st quartile (Q₁), 2^nd quartile (Q₂), 3^rd Quartile (Q₃), and 4^th quartile (Q₄) of total maize yield, respectively.