Complete the class below which implements a KNN classifier $($and meets the requirements for being a sklearn predictor$).$ You only

need to complete the fit$()$ and predict$()$ methods. You may modify the contents of any methods as you see fit $($but not the

signatures$),$ and you may make any other functions you want $($but prefix any additional functions with an underscore to avoid conflicts

with sklearn$).$

Details:

$K$ will always be less than the number of training points.

All features will be numeric.

The number of features will be $2.$

Labels can be anything hashable $($so they can go in a dict or set$).$

Use euclidean distance to compute distance between points.

In the event of a tie, you may choose which of the tied labels to predict.

predict$()$ must return a numpy.ndarray.

You may not use any sklearn functionality aside from the uses already provided $($ sklearn. base. BaseEstimator and

sklearn.metrics.accuracy$\_$score$).$class MyKNN$($sklearn$.$base.BaseEstimator$)$: