Fig. 4.22. The ELR model constructed for the olive oil content data.

e 4.23 shows the coefficient profiles of the RLR model, the Lasso

d the ELR model constructed for the olive oil content data. Their

ߣ values are also displayed. Using different constrained regression

he optimal ߣ values were different. It can be seen that the optimal

was larger in the RLR model, was smaller in the Lasso model,

n between in the ELR model.

(a) (b) (c)

The constrained regression models constructed for the oil content data. (a) The

l. (b) The Lasso model. (c) The ELR model.

king variables using the vip package

package can also be used to evaluate the importance of variables

the RLR, Lasso and ELR models as well as other models shown