Consider the following training dataset. The features are F1, F2 and F3, and the class label is C. We want to construct the decision tree. Using the gain ratio, determine whether F1 or F3 is the best attribute to start splitting on. Justify your answer by indicating the value of the gain ratio for F1 and F3. F2 F3 C F1 X M C1 IA | C1 B Y 0 C1 B M C1 B IN C2 C Y IN C2 C Y 0 C2 C X IN C2 D | C1 D IN C2