The following is a training dataset to be used by a k$-$NN classifier.

$\backslash$table$[[$Fo$.,$Clatures,$\backslash$table$[[$Class$],[$Label$]]],[,$Feature $1,$Feature $2,$Feature $3,$Feature $4,$Feature $5,$Feature $6,,,,,,,,],[1,1,3,4,5,1800,4,1,,,,,,,],[2,2,3,3,4,1900,3,1,,,,,,,],[3,4,2,2,3,1700,3,1,,,,,,,],[4,10,12,14,10,2500,11,2,,,,,,,],[5,11,12,11,13,2400,14,2,,,,,,,],[6,12,12,14,14,2300,12,2,,,,,,,],[7,13,14,12,12,2400,13,2,,,,,,,],[8,14,14,13,13,2600,12,2,,,,,,,],[9,12,11,13,12,2650,11,2,,,,,,,],[10,13,14,12,14,2600,12,2,,,,,,,]]$

You are given the following test sample:

$\backslash$table$[[$Feature $1,$Feature $2,$Feature $3,$Feature $4,$Feature $5,$Feature $6],[2,2,4,3,4000,2]]$

a$)$ Using Euclidian distance and $k=6$ classify the test sample $($show all your calculations$).$ What is the problem with this choice of $k.$

b$)$ In a $2-$class problem solved by using a k$-$NN classifier, what would be the challenge with setting $k=2?$