无法在<module'__main __'>上获取属性'MyClassifier'

Question

我目前正在研究音频，以按情感对它们进行分类。我使用的策略是将每种情感与其他情感进行分类。为了开心，我使用了这段代码并取得了商品结果，然后保存了模型：

from sklearn.base import BaseEstimator

class MyClassifier(BaseEstimator):

    def __init__(self, classifier_type: str = 'XGBoost'):  #""A Custome BaseEstimator that can switch between
        #classifiers :param classifier_type: string - The switch for different classifiers



        self.classifier_type = classifier_type


    def fit(self, X, y=None):
        if self.classifier_type == 'XGBoost':
            self.classifier_ = XGBClassifier()

        elif self.classifier_type == 'RandomForest':
            self.classifier_ = RandomForestClassifier()


        elif self.classifier_type == 'CART':
            self.classifier_ = DecisionTreeClassifier()

        elif self.classifier_type == 'KNearestNeighbord':
            self.classifier_ = KNeighborsClassifier()

        elif self.classifier_type == 'GaussianNB':
            self.classifier_ =  GaussianNB()

        elif self.classifier_type == 'SupportVectorMachine':
            self.classifier_ = SVC(kernel = "linear")

        elif self.classifier_type == 'LogisticRegression':
            self.classifier_ = LogisticRegression()


        elif self.classifier_type == 'SGDClassifier':
            self.classifier_ = SGDClassifier()   

        else:
            raise ValueError('Unkown classifier type.')

        self.classifier_.fit(X, y)
        return self

    def predict(self, X, y=None):
        return self.classifier_.predict(X)

#feature selection
clf = ExtraTreesClassifier()
clf = clf.fit(X, y)
clf.feature_importances_  
model = SelectFromModel(clf, prefit=True)
X_selected = model.transform(X) 


#split train and test
training_size=0.7 #70% of the data set for training

testsize= 1-training_size
seed=30 
X_train,X_test,y_train,y_test=train_test_split(X_selected, y,
                                               train_size= training_size, random_state=seed, test_size=testsize)


pipeline = Pipeline([
    ('scale', StandardScaler()),
    ('clf', MyClassifier())
])

parameter_space = {
    'clf__classifier_type': ['XGBoost', 'RandomForest',
  'CART', 'KNearestNeighbord', 
                             'GaussianNB', 'SupportVectorMachine', 'LogisticRegression', 'SGDClassifier']
}

from sklearn.model_selection import GridSearchCV
prec_scorer = make_scorer(precision_score, average="micro")

search = GridSearchCV(pipeline , parameter_space, n_jobs=-1, cv=5, scoring= prec_scorer)
search.fit(X_train, y_train)

print('Best model:\n', search.best_params_, search.best_score_)
#save the model to disk
filename = 'happy_MODEL.pkl'
pickle.dump(search.best_estimator_, open(filename, 'wb'))

但是问题是，当我使用保存的模型预测新数据的准确性时，我遇到了这个错误：'无法获取属性'MyClassifier' '。我不理解这个问题。为了测试我的模型，我使用以下代码：

filename = 'happy_MODEL.pkl'
loaded_model = pickle.load(open(filename, 'rb'))
result = loaded_model.predict(X_selected)
print('\n', accuracy_score( y, result))

有人可以帮我吗？