Understanding the impact of 'k'

There is a complex relationship between k and classification accuracy. Bigger is not always better.

Which of these is a valid reason for keeping k as small as possible (but no smaller)?

Answer the question

50 XP

Possible Answers

A smaller k requires less processing power
press1
A smaller k reduces the impact of noisy data
press2
A smaller k minimizes the chance of a tie vote
press3
A smaller k may utilize more subtle patterns
press4