A genetic program is used to evolve a model tree. Which of the following will be the best for the fitness function?

Question $8$Select one:

a$.$

Calculate the entropy for each leaf node, and minimize the average entropy over all leaf nodes.

b$.$

For each leaf node, compute the standard deviation over the target values. Compute the average deviation over all of the leaf nodes. Minimize this standard deviation.

c$.$

Calculate the entropy for each leaf node, and minimize the average entropy over all leaf nodes. Add to this a tree complexity, where the tree complexity is the depth of the tree.

d$.$

For each leaf node, compute the standard deviation over the target values. Compute the average deviation over all of the leaf nodes. Add to the average standard deviation a tree complexity penalty, where the tree complexity is depth of the tree. Minimize this penalized fitness function.