Question



Output of the KNN algorithm question -> Prediction or probability

Hello Dr. Ballings,

In the question on KNN algorithm (2.8.4), you asked us to -
Compute the proportions of 1s in churn for all of "df_new", in other words, make a prediction for all instances in "df_new"

So, should our output be the probability of a 1 in churn (i.e. a number between 0 and 1) or a concrete prediction (i.e. 0 or 1)?

My initial thought is since you said prediction it should be 0 or 1, but since you also asked for computing the proportions, I got a little
confused.

-Thanks,
Tapajit Dey





Answers and follow-up questions





Answer or follow-up question 1

Dear Tapajit,

It should be a score (probability) between 0 and 1.

Michel Ballings





Sign in to be able to add an answer or mark this question as resolved.