In one exam you asked what would happen if we took an algorithm that learns rectangles and multiplied the data by a scalar. The answers say that the result wouldn't change because the classifier's boundaries would also be multiplied by that same scalar. Isn't that the definition of the result changing? In these types of questions, what do you mean by "will the result change?".

Also, what is expected of us when you ask "For a soft-margin SVM with a polynomial kernel, what will happen if we multiply X by A (s.t. A is of rank n)?" It seems a little hard to visualize…

Thanks again.