Consider three classification models discussed in class: K$-$Nearest Neighbor, Logistic Regression,

and Decision Trees. Suppose we have some binary outcome data with $2$ features, x$1$ and x$2.$

While all the algorithms can handle types of data, they can be better or worse at some datasets.

$($a$)$ For each of the $3$ algorithms, make a scatter plot $($x$1$ vs$.$ x$2)$ of a potential dataset where

that algorithm would do perhaps struggle some, and explain why. When plotting, you can

use different shapes or colors for the $1$ or $0$ outcomes.

$($b$)$ For each of the $3$ algorithms, explain how important it is to normalize the data before

training.