2. Linear Regression Learner (40 points) 1) Split the voting dataset into training and testing subsets by the ratio 80/20. a. Train a Logistic Regression classifier, what is the prediction accuracy of the model? b. Print the probability estimates for the prediction. Compared to decision tree, is there any difference? c. Can we use Linear Regression to model any of the three problems? Why? 3. (*) Regression Analysis (30) noints) The above table lists the 14 days of Jason playing tennis records. Use NumPy to represent the data as array 1) Train a linear regression model with the above dataset, what is R-squared score? What is the prediction result for Day 15 (Rain, Hot, High, Weak)? 2) Train a regression tree model with the above dataset, what are the height and number of leaves? What is the prediction result for Day 15 (Rain, Hot, High, Weak)? 3) In decision tree learning, we use information entropy as the splitting criterion, why not use loss or error (like the loss used in linear regression) to split nodes, i.e., always choose the feature with the minimum loss for subsets after splitting the set