function [label_test] = knn(k, data_train, label_train, data_test)% knn - k nearest neighbours classifiererror(nargchk(4,4,nargin));dist = l2_distance(data_train, data_test);[sorted_dist, nearest] = sort(dist);nearest = nearest(1:k,:);label_test = label_train(nearest);% note this only works for binary labelslabel_test = mean(label_test,1) >= 0.5;