Instantiate the underlying CatBoostClassifier model

Parameters ---------- train_sample contains the following torch.Tensors: `(past_target, past_covariates, historic

`train_sample` contains the following tensors: (past_target, past_covariates, historic_future_covariates, future_covariates, stat

Returns the decoder variable importance of the TFT model. The decoder_weights are calculated for the known future inputs of the model.

applies cyclic encoding from `datetime_attribute_timeseries()` to `self.attribute` of `index`.

Encode `index` as a scalar.

Apply the user-defined callable to encode the index.

Returns the encoder variable importance of the TFT model. The encoder_weights are calculated for the past inputs of the model. The en

Helper function, used to make FFT model pickable.

Returns whether all the transformers in the pipeline were fitted (when applicable). Returns ------- bool

The training values are considered as the scorer model

Train one sub-model for each component when self.is_univariate=False and series is multivariate

Fit the filters (if applicable) and scorers.

Fit the forecasting model (if applicable) and scorers.

Custom fit function for `SKLearnModelWithCategoricalFeatures` models, adding logic to let the model handle categorical features direc

Validate and format the categorical columns listed in self._categorical_indices accordingly to the model's requirements.

For some reason CatBoostClassifier does regression when given continuous labels For consistency, an error is artificially raised on c

Wrapper function to adapt the SHAP explainer to the torch forecasting model. It takes as input a numpy array of shape `(num_samples,

This model is recurrent, so we have to write a specific way to obtain the time series forecasts of length n.

Builds the kwargs dictionary for the transformation function. Parameters ---------- transformation

Gets the class label with highest predicted probability per component extracted from the model output.

Sanity checks for the historical_forecasts function Parameters ---------- args The args parameter(s) provided to