time_series_predictor package

Subpackages

Submodules

time_series_predictor.l1_regularized_nnr module

L1 regularized NNR

https://skorch.readthedocs.io/en/stable/user/neuralnet.html?highlight=NeuralNet%20get_loss#subclassing-neuralnet

class time_series_predictor.l1_regularized_nnr.L1RegularizedNNR(*args, lambda1=0, **kwargs)

Bases: NeuralNetRegressor

L1 regularization

\[L_{loss}=\left \| y-\hat{y} \right \|^2+\lambda | W |\]

L1 regularization makes the weight vector sparse during the optimization process.

The optimizer in PyTorch can only implement L2 regularization, and L1 regularization needs to be implemented manually, that is the purpose of this class.

Note

This example also regularizes the biases, which you typically don’t need to do.

get_loss(y_pred, y_true, X=None, training=False)

Return the loss for this batch.

Parameters

  • y_pred (torch tensor) – Predicted target values

  • y_true (torch tensor) – True target values.

  • X (input data, compatible with skorch.dataset.Dataset) –

    By default, you should be able to pass:

    • numpy arrays

    • torch tensors

    • pandas DataFrame or Series

    • scipy sparse CSR matrices

    • a dictionary of the former three

    • a list/tuple of the former three

    • a Dataset

    If this doesn’t work with your data, you have to pass a Dataset that can deal with the data.

  • training (bool (default=False)) – Whether train mode should be used or not.

set_input_shape(X, y)

time_series_predictor.min_max_scaler module

min_max_scaler

class time_series_predictor.min_max_scaler.MinMaxScaler

Bases: BaseEstimator, TransformerMixin

Min Max Scaler class

fit(X, *_args, **_kwargs)

Compute the minimum and maximum to be used for later scaling.

Parameters

  • input_matrix – input matrix

  • axis – None or int or tuple of ints, optional Axis or axes along which to operate.

fit_transform(X, y=None, **fit_params)

Fit to data, then transform it.

Parameters

input_matrix – input matrix

Returns

transformed matrix

inverse_transform(transformed)
Parameters

transformed – transformed input

Returns

inverse transformed

transform(X)

Scale features of input_matrix according to feature_range.

Parameters

input_matrix – input matrix

Returns

transformed matrix

time_series_predictor.time_series_predictor module

time_series_predictor script

class time_series_predictor.time_series_predictor.CheckpointHandler

Bases: Callback

on_train_end(net, X=None, y=None, **kwargs)

Called at the end of training.

class time_series_predictor.time_series_predictor.InputShapeSetter

Bases: Callback

dynamically set the input size of the PyTorch module based on the data

Typically, it’s up to the user to determine the shape of the input data when defining the PyTorch module. This can sometimes be inconvenient, e.g. when the shape is only known at runtime. E.g., when using sklearn.feature_selection.VarianceThreshold, you cannot know the number of features in advance. The best solution would be to set the input size dynamically.

on_train_begin(net, X=None, y=None, **kwargs)

Called at the beginning of training.

class time_series_predictor.time_series_predictor.ScoreCalculator(scores_to_calculate, locks, score_method, output_list)

Bases: Thread

Spread computational effort across cpus

run()

Method representing the thread’s activity.

You may override this method in a subclass. The standard run() method invokes the callable object passed to the object’s constructor as the target argument, if any, with sequential and keyword arguments taken from the args and kwargs arguments, respectively.

class time_series_predictor.time_series_predictor.TimeSeriesPredictor(net, early_stopping=None, **l1_regularized_nnr_params)

Bases: object

Network agnostic time series predictor class

Parameters

  • **l1_regularized_nnr_params (skorch L1RegularizedNNR parameters.) –

  • early_stopping (torch.callbacks.EarlyStopping object) –

fit(dataset, **fit_params)

Fit selected network

Parameters

  • dataset (dataset to fit on) –

  • net (network to use) –

  • **fit_params (dict) – Additional parameters passed to the forward method of the module and to the self.train_split call.

forecast(*args, **kwargs)

Future forecast

Parameters

  • *args (variable length unnamed args list) –

  • **kwargs (variable length named args list) –

Returns

future forecast

Returns

future dataframe

make_future_dataframe(*args, **kwargs)
Parameters

  • *args (variable length unnamed args list) –

  • **kwargs (variable length named args list) –

Returns

future dataframe

predict(inp)

Run predictions

Parameters

inp – input

sample_forecast(*args, **kwargs)

Future forecast

Parameters

  • *args (variable length unnamed args list) –

  • **kwargs (variable length named args list) –

Returns

future forecast

Returns

future dataframe

sample_predict(inp)

Run predictions

Parameters

inp – input

score(dataset)

Compute the mean r2_score of a network on a given dataset.

Parameters

dataset – dataset to evaluate.

Returns

mean r2_score.

Module contents

__init__.py