(iii) Lets now generate a new dataset that includes the 2 attributes X and Y from the original data set, and 95 exact duplicates of X. What impact will this have on the relative suitability of all three classifiers mentioned above? Which one is best? Which one is the worst? Briefly explain.

Impact on 1-NN:

Impact on Decision Tree:

Impact on Perceptron:

Best performing classifier:

Worst performing classifier:

(iv) We generate another new dataset, which includes the 2 attributes X, and Y from the original data set, and 95 additional noisy attributes. The values for the noisy attributes are randomly assigned from the uniform distribution on [0, 20]. What impact will this have on the performance of each of these 3 classifiers? Briefly explain.

Impact on 1-NN:

Impact on Decision Tree:

Impact on Perceptron:

18 16 + 14 12 10 8 6 6 4 2 0 2 6 8 10 12 16 18 20 X 18 16 + 14 12 10 8 6 6 4 2 0 2 6 8 10 12 16 18 20 X