Matlab2016b版本代碼：

class = knnclassify(X_test, X_train, Y_train,k); % [X_train,X_test,Y_train, Y_test],X樣本，Y標簽，train訓(xùn)練，test測試

%k可以為1,2,3,4.....。X_test為測試數(shù)據(jù)，X_train，Y_train為訓(xùn)練集的樣本和標簽，k為設(shè)置的k臨近值的k值

Matlab2019b對應(yīng)的代碼為：

mdl = fitcknn(X_train, Y_train,'NumNeighbors',k);? ? %k為對應(yīng)的1,2,3,4.....

class = predict(mdl,X_test);