What is a disadvantage of 1-NN? How can it be overcome?
1-NN poor generalization
-> use k-NN and majority vote to be robust against outliers / errors in training dataset
Zuletzt geändertvor 2 Jahren