2. (16 points) Decision trees. (a) (8 points) For each of the following data sets, explain whether or not a basic decision tree of depth 2 will excel in classifying the data. If not, propose a classifier that will. [CI ID] (b) (4 points) Consider training AdaBoost on the following data set, where the data are linearly separable except for the existence of an outlier. What would happen to the weight assigned to that outlier after many boosting iterations? Would you conclude that AdaBoost is robust to outliers? 05 D 020 (c) (4 points) In AdaBoost, would you stop the iteration if the error rate of the current weak learner on the weighted training data is 0? Explain. 2. (16 points) Decision trees. (a) (8 points) For each of the following data sets, explain whether or not a basic decision tree of depth 2 will excel in classifying the data. If not, propose a classifier that will. [CI ID] (b) (4 points) Consider training AdaBoost on the following data set, where the data are linearly separable except for the existence of an outlier. What would happen to the weight assigned to that outlier after many boosting iterations? Would you conclude that AdaBoost is robust to outliers? 05 D 020 (c) (4 points) In AdaBoost, would you stop the iteration if the error rate of the current weak learner on the weighted training data is 0? Explain