10주차 복습 : KNN, SVM, Ensemble
##위에서 찾은 다른 best parameter들은 쓰고, n_neighbors 하이퍼파라미터만 조정해서 정확도 찾았음 train_acc = []; test_acc = [] candidates = [1,2,3,4,5] for cand in candidates: knn = KNeighborsClassifier(metric = 'euclidean', n_neighbors = cand, weights = 'uniform') knn.fit(X_train, y_train) train_acc.append(knn.score(X_train, y_train)) pred = knn.predict(X_test) test_acc.append(accuracy_score(y_test, pred)) fig = plt.figure(..
2020. 12. 11.
