Answer Happy

Question 7 (Mandatory) (1 point) Saved Fred has a binary classification task with n=72 observations each measured on p=7125 items of real value. One of the following techniques is inherently not designed to handle such a structure underlying the classification task logistic regression support vector machine with linear kernel support vector machine with Gaussian RBF Kernel naive bayes classifier
Question 8 (Mandatory) (1 point) With a properly chosen margin, and thanks to the large margin principle, the average test error of an SVM classifier should be lower than the Bayes Risk True False Question 9 (Mandatory) (1 point) The dual implementation of the Support Vector Machine classifier gets considerably slow when the sample grows larger while p remains fixed and small True False Saved