medium10 pts · first answer countsmachine-learningknnbias-variance

kNN with k = 1

A k-nearest neighbours classifier uses k = 1 and is scored on its own training data (no duplicate points with different labels). What do you expect?

A.Low training accuracy and high biasB.The same accuracy on train and testC.Training accuracy near 50% on balanced dataD.Training accuracy of 100% and high variance on new data

Discussion 0

Join free to post in the discussion.

kNN with k = 1

Discussion 0

Related questions