QUESTION 7 Select all valid options Storing only points whose labels are incorrectly predicted, rather than all training instances, reduces time to classify test instances but increases training time. Increasing k will necessarily increase the classification accuracy of kNN on unseen examples. It is preferable to use an even, rather than odd, value for k. The predictive model learned by KNN, when storing only points whose labels are incorrectly predicted, is independent of the sequence in which the input instances are considered