Как мне использовать smote syntheti c меньшинства в python?

Question

Я делаю модель KNN, и у меня очень несбалансированные данные, такие как:

Buy     515     
Don't move  1473
Sell    466

Теперь я хочу перевыбор метки покупки и продажи, чтобы она равнялась количеству не двигаться в моей модели KNN.

Это моя модель KNN, но я не знаю, как сделать SMOTE.

knn_neighbours = [2,3,4,5,10]
knn_weights = ['uniform','distance']
knn_leaf = [10,20,30,40,80]

knn_grid_search = list(product(knn_neighbours,knn_weights,knn_leaf))
knn_grid_results = {}

for option in knn_grid_search:
    knn_neighbours = option[0]
    knn_weights =  option[1]
    knn_leaf = option[2]

    X_data = result
    y_data = y


    knn_macro_list = []
    knn_micro_list = []
    knn_auc_list = []
    acc_knn_score1 = []


    from sklearn.model_selection import StratifiedKFold
    skf = StratifiedKFold(n_splits=10, shuffle=True)
    for train_index, test_index in skf.split(result, y):


train_news, test_news = result.iloc[train_index], result.iloc[test_index]
            train_labels, test_labels = y[train_index], y[test_index]

        le = preprocessing.LabelEncoder()
        le.fit(train_labels)
        classes_name=le.classes_
        le.transform(train_labels)
# Binarize the output
        train_labels = le.transform(train_labels)
        test_labels = le.transform(test_labels)

        train_labels = label_binarize(train_labels, classes=[0,1,2,3,4])
        test_labels = label_binarize(test_labels, classes=[0,1,2,3,4])

        n_classes = train_labels.shape[1]

#K-Nearest Neighbour

     knn = KNeighborsClassifier(n_neighbors=knn_neighbours,weights=knn_weights,leaf_size=knn_leaf)


acc_knn_score1.append(knn.fit(train_news,train_labels).score(test_news, test_labels))

Научиться прогнозировать каждый класс против другого

        knn_classifier = OneVsRestClassifier(knn)

knn_score = knn_classifier.fit(train_news, train_labels).predict_proba(test_news)
        acc_knn_score = knn_classifier.score(test_news, test_labels)