AttributeError: 'numpy.ndarray' object has no attribute 'fit' when calling fit_transform on a pipeline

Question

I'm getting the below error when I call pipeline.fit_transform(X_train, y_train).

AttributeError: 'numpy.ndarray' object has no attribute 'fit'

The individual transformers in the pipeline work fine, but when I combine them in the pipeline I get the error.


X, y = training_data.drop('Response', axis=1), training_data['Response']
X_train, X_test, y_train, y_test = train_test_split(X, y, stratify=y)

X_train = np.array(X_train)
X_test = np.array(X_test)
y_train = np.array(y_train)
y_test = np.array(y_test)

class preprocess(TransformerMixin, BaseEstimator):

    def __init__():
        self.X = None

    def fit(self, X, y=None):
        self.X = X
        self.PI2 = 'Product_Info_2'
        self.PI2_categories = list(training_data[self.PI2].unique())
        return self

    def transform(self, X, y=None):
        Xt = X.copy()
        Xt = pd.concat([Xt, pd.get_dummies(Xt[self.PI2])], axis=1).drop(self.PI2, axis=1)
        Xt.drop('Id', axis=1, inplace=True)
        Xt.fillna(value=0, inplace=True)
        return np.array(Xt)


class apply_NB(TransformerMixin, BaseEstimator):

    def __init__(self):
        self.gridCV = None
        self.params = {"var_smoothing": [x*10**(-9) for x in [0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 
                                                              0.8, 0.9, 1, 1.5, 2, 2.5, 3, 3.5, 
                                                              4, 4.5, 5]]}
        self.best_params = None

    def fit(self, X, y):

        self.gridCV = GridSearchCV(GaussianNB(), self.params, verbose=10, n_jobs=-1)
        self.gridCV.fit(X, y)
        self.best_params = self.gridCV.best_params_
        return self

    def transform(self, X, y=None):
        Xt = self.gridCV.predict(X)
        return Xt

nb_pipeline = Pipeline([('preprocess', preprocess),
                        ('fit_NB', apply_NB())])

nb_pipeline.fit_transform(X_train, y_train)

When I try the final line I just get:

AttributeError: 'numpy.ndarray' object has no attribute 'fit'

pkapka · Accepted Answer · 2019-05-12 22:21:04Z

4

You forgot to put self in the first init of preprocess

class preprocess(TransformerMixin, BaseEstimator):

    def __init__(self):
        self.X = None

and then you gotta initialize the class for applyNB too.

nb_pipeline = [('preprocess', preprocess()),
                    ('fit_NB', apply_NB())]

Seems to work for m after making these changes!

answered May 12, 2019 at 22:21

pkapka

5,3761 gold badge19 silver badges22 bronze badges

Sign up to request clarification or add additional context in comments.

Collectives™ on Stack Overflow

AttributeError: 'numpy.ndarray' object has no attribute 'fit' when calling fit_transform on a pipeline

1 Answer 1

Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

Comments

Your Answer

Sign up or log in

Post as a guest

Related