Q3 (8pts): Return of the Wine dataset that has been processed to be all Categorical data type: Q3a: List all the possible splits at the root, using only the Alcohol feature. *You are NOT expected to fill all the rows List which training samples would end up in each leaf based on that split Q3b: Your goal is building a classification tree for predicting quality. Calculate the TWGI impurity of the following two splits. Color==Red Citric Acid == Medium Q3c: Which split would you pick? Q3d: What Quality would you predict for sample 10 if you use the Color == Red split for the root and stopped building the tree at this point