Question

function predictKNN which computes the KNN prediction for given input xi, K and train data.

A correct implementation will behave similar to below example:

Input: xi: list, trainx: list of lists, trainy: list of list, k: integer

call: predictKNN(trainx[0],trainx,trainy)

output: 2: integer

def predictKNN(xi,trainx,trainy,k=3):
###
### YOUR CODE HERE
###

Answer 1

Assume that you are given list_x , list_y, list_target where list_x has x values,  list_y has corresponding y values and list_target has the corresponding target values where the target is 0 or 1.

Assuming this, here is the python code:

import pandas as pd
import numpy as np
from sklearn.preprocessing import StandardScaler
from sklearn.neighbors import KNeighborsClassifier

def KNN(listx,listy,listtarget,testOrdPair,k):
df = pd.DataFrame(list(zip(listx,listy,listtarget)),columns=['x','y','target'])
scaler = StandardScaler()
scaler.fit(df.drop('target',axis=1))
scaled_features=scaler.transform(df.drop('target',axis=1))
x=df_feat=pd.DataFrame(scaled_features,columns=df.columns[:-1])
y=pd.DataFrame(df['target'])
knn = KNeighborsClassifier(n_neighbors=k)
knn.fit(x,y)
pred=knn.predict(testOrdPair)
return pred

Here. testOrdPair refers to a list of (x,y) for which you want a prediction, a list of targets corresponding to each ordered pairs is returned.

Hope this helps!