If the purpose is to predict MEDV for several thousands of new tracts, what would be the

disadvantage of using k$-$NN prediction? List the operations that the algorithm goes through to

produce each prediction.