$2\mathrm{.}2.$ sample run $3$

What is the patient's ID: $200$

Majority vote from $5$ nearest neighbors: Positive

Actually this patient is tested: Negative. Step $2$: Run over all patients and check the overall accuracy for $\backslash (\backslash$mathrm$\{$k$\}=10\backslash )$

Based on what you have in step $1,$ you need to run a loop to check every patient's prediction vs actual diagnosis. Then find the accuracy $($total number of correct predictions$/$total cases$)$

$\backslash$# $(5$ points$)$ for any given $\backslash ($ k $\backslash ),$ for example $\backslash ($ k$=10\backslash ),$

$\backslash$# loop over all patients then find the accuracy Step $3$: Find the best k$.$

For different K$,$ this accuracy might not be the same. Find the best K which corresponds the highest accuracy.

$16]$: $\backslash$# $(5$ points$)$ loop for different k: from $1$ to $50($this loop will take $3-5$ minutes$)$

$\backslash$# find the best $\backslash ($ k $\backslash )$ which corresponds the highest accuracy

Best k: $5$ The max. accuracy $0\mathrm{.}8283828382838284$

$3\mathrm{.}2.$ plotting $($optional for you$)$

$17]$: $\backslash$# $(0$ points$)$ plotting $3\mathrm{.}2.$ plotting $($optional for you$)$

$(5$ points$)$ comparison: with$/$without data scaling:

$-$ rerun your code if you skipped the step of data scaling in the beginning.

$-$ what are the differences?