Exercise 8 (Genetic algorithms and decision trees)

Let's suppose we have the same data sets as in tasks 1 and 2. The data sets are so small that we cannot learn decision trees accurately. However, we could use genetic algorithms to generate an optimal decision tree from the data.

Describe how genetic algorithms can be used for optimizing a decision tree! Describe carefully the coding of solutions (decision trees), initial population, cross-over and mutation operations and the selection criterion.