Section $3$

Dana was given a hard classification problem and she decided to use SVM with polynomial kernel with $d=2,10,20.$ For each degree, she tried $15$ to $85$ training samples, with jumps of $.$