(a) (b)

a) The importance visualisation of the variable ranking result of the SVM

he olive oil content data. The model employs ten hidden neurons. (b) The fitness

nt for the SVM model.

andom forest regression algorithm can also help ranking variables

igure 4.27(a) shows the variable importance ranking result of the

orest model constructed for the olive oil content data. The ranking

done using the vip package. In this model, the paste water

was ranked at the top. Figure 4.27(b) shows the fitness

ments for this model. The R-square of the model was 0.654 and

tistic p value of the model was 6.05e−6.

he random forest model for the oil content data. (a) The importance measure of

b) The fitness measurements of the model.

e 4.28 shows a decision tree generated using the party package.

node shown in Figure 4.28 employed the fruit weight variable as

partitioning rule. Using this tree, it can be seen that maximising

ontent in olives must go through two partitioning rules, i.e., the