Now that we have have our data, we need to split this into a training set, and a testing set. But before we split our data into training and testing, we also need to split our data into the predictive features $($denoted $x)$ and the response $($denoted $y).$

Write a function that will take as input a $2-$d numpy array and retum four variables in the form of $($X$\_$train, y$\_$train$),($X$\_$test, $y_{-}$test$),$ where $($X$\_$train, y$\_$train$)$ are the features $+$ response of the training set, and $(x-$test, $y_{-}$test$)$ are the features $+$ response of the testing set.

Function Specifications:

Should take a $2-$d numpy as input.

Should split the array such that $x$ is the year, and $y$ is the corresponding population.

Should return two tuples of the form $($X$\_$train, $y_{-}t$ rain$),($X$\_$test, $y_{-}$test$).$

Should use sklearn's train$\_$test$\_$split function with a test$\_$size $=0\mathrm{.}2$ and random$\_$state $=42.$

### START FUNCTION

def feature$\_$response$\_$split$($arr$)$:

# your code here

return