Question 1:

The K nearest neighbor algorithm has hyper-

parameters in addition to K. Use the titanic dataset

to simultaneously tune the values of more than one

hyper-parameter. E.g. Value of K, weight, and

metric. You can either do nested for loops for this or

use the grid search OR random search function

from Scikit-learn. (Please submit code inside of a

Jupyter notebook). Based on your analysis, propose

a strategy for selecting an optimal value of k for a

given dataset and discuss its advantages and

limitations.

10 points

Describe and analyze the impact of varying the

value of k in the KNN classifier on its classification

performance considering both accuracy metrics.

5 points