def fit(self, X, y, input_checks=True, validation_X=None,
validation_y=None, **kwargs):
"""
Fit the regressor on the training set (X, y)
----------
X : a nested pd.Dataframe, or (if input_checks=False) array-like of
shape = (n_instances, series_length, n_dimensions)
The training input samples. If a 2D array-like is passed,
n_dimensions is assumed to be 1.
y : array-like, shape = [n_instances]
The training data class labels.
input_checks : boolean
whether to check the X and y parameters
validation_X : a nested pd.Dataframe, or array-like of shape =
(n_instances, series_length, n_dimensions)
The validation samples. If a 2D array-like is passed,
n_dimensions is assumed to be 1.
Unless strictly defined by the user via callbacks (such as
EarlyStopping), the presence or state of the validation
data does not alter training in any way. Predictions at each epoch
are stored in the model's fit history.
validation_y : array-like, shape = [n_instances]
The validation class labels.
Returns
-------
self : object
"""
X = check_and_clean_data(X, y, input_checks=input_checks)
validation_data = \
check_and_clean_validation_data(validation_X, validation_y)
# ignore the number of instances, X.shape[0],
# just want the shape of each instance
self.input_shape = X.shape[1:]
if self.batch_size is None:
self.batch_size = int(max(1, min(X.shape[0] / 10, 16)))
else:
self.batch_size = self.batch_size
self.model = self.build_model(self.input_shape)
if self.verbose:
self.model.summary()
self.history = self.model.fit(
X,
y,
batch_size=self.batch_size,
epochs=self.nb_epochs,
verbose=self.verbose,
callbacks=self.callbacks,
validation_data=validation_data,
)
self.save_trained_model()
self._is_fitted = True
return self
def fit(self, X, y, input_checks=True, validation_X=None,
validation_y=None, **kwargs):
"""
Fit the classifier on the training set (X, y)
----------
X : a nested pd.Dataframe, or (if input_checks=False) array-like of
shape = (n_instances, series_length, n_dimensions)
The training input samples. If a 2D array-like is passed,
n_dimensions is assumed to be 1.
y : array-like, shape = [n_instances]
The training data class labels.
input_checks : boolean
whether to check the X and y parameters
validation_X : a nested pd.Dataframe, or array-like of shape =
(n_instances, series_length, n_dimensions)
The validation samples. If a 2D array-like is passed,
n_dimensions is assumed to be 1.
Unless strictly defined by the user via callbacks (such as
EarlyStopping), the presence or state of the validation
data does not alter training in any way. Predictions at each epoch
are stored in the model's fit history.
validation_y : array-like, shape = [n_instances]
The validation class labels.
Returns
-------
self : object
"""
if self.callbacks is None:
self.callbacks = []
self.random_state = check_random_state(self.random_state)
X = check_and_clean_data(X, y, input_checks=input_checks)
y_onehot = self.convert_y(y)
# ignore the number of instances, X.shape[0],
# just want the shape of each instance
self.input_shape = X.shape[1:]
self.nb_classes = y_onehot.shape[1]
self.adjust_parameters(X)
X, y_onehot, _ = self.pre_processing(X, y_onehot)
validation_data = \
check_and_clean_validation_data(validation_X, validation_y,
self.label_encoder,
self.onehot_encoder)
if validation_data is not None:
vX, vy, _ = self.pre_processing(validation_data[0],
validation_data[1])
validation_data = (vX, vy)
input_shape = X.shape[1:]
self.model = self.build_model(input_shape, self.nb_classes)
self.history = self.model.fit(
X,
y_onehot,
batch_size=self.batch_size,
epochs=self.nb_epochs,
verbose=self.verbose,
callbacks=self.callbacks,
validation_data=validation_data,
)
self.save_trained_model()
self._is_fitted = True
return self
