The algorithm for KNN classification model is given below:

Algorithm: KNearestNeighbors

$,$Input: $,$ Training data $x$

$,,$Training

$,,$data labels

$,,Y$ Sample $x$

$,,$to classify

$,$Output: Decision yp about sample $x$

$,$for ilarr$1$ to $m$ do

$,,$Compute distance between training sample $$ and

$,,$unlabeled sample $x$ i$.$e$.d(,x)$

$,$end for

Compute set I containing the indices for the $k$ smallest distances $d(x_i,x)$ Compute the decision class $y$ pby measuring the majority

label Y from I return $y$ p

Task $1$

Write a program to implement KNN classifier on heart$\_$statlog$\_$cleveland$\_$hungary$\_$data and classify given dataset.

Load dataset

Perform required preprocessing on data

Split dataset into test and train sets

Perform $KNN$ algorithm to make predictions for $k=5$ and $(6-10)$ as well.

Compute accuracy and confusion matrix

Find accuracy for each value of $k$ and display and for $3$ different splits as $30$:$70,20$:$80,10$:$90$

Find the best $k$ which gives highest value of accuracy and worst $k$ which gives worst accuracy

Also compute confusion matrix for both best and worst value of $k.$