Q3 [2.5]. Classifying the iris data set. In this question, you will experiment with different classification models on the Iris data set. The models are: Linear Classifiers, with hinge loss (SVM) Random-forest Classifier

a) for the SVM classifier, use the default parameters, and 5-fold cross-validation, and report the overall accuracy and confusion matrix, as well as accuracy and confusion matrix for each fold. What is the standard deviation of accuracy over the folds?

b) for the Random Forest classifier, experiment with different numbers of estimators, different max depths of the tree (in the range from 2 to 10). Use the default values for the remaining parameters. Evaluate each parameter selection using 5-fold cross-validation, and report the accuracy and the confusion matrix. Only report the interesting parameter settings.

c) Summarize your findings from (a) and (b). What is the best performing classifier on the iris data set?