in knn classification, how are the predictions made? by averaging the target values of the k nearest neighbors, by taking the majority vote of the k nearest neighbors, by fitting a polynomial curve to the k nearest neighbors, by calculating the median of the target values of the k nearest neighbors