Defines (in)equality between two likelihood objects.

Return a new series with elements selected by `key`. The supported index types are the following base types as a single value, a list or a sl

Gets subset of Pipeline based either on index or slice with indexes. Resulting pipeline will deep copy transformers of the original p

Returns a sample drawn from this dataset.

Returns a sample drawn from this dataset.

Returns a sample drawn from this dataset.

Manager for all Darts configuration options.

Create a ``TimeSeries`` from a time index `times` and values `values`. See Also -------- TimeSeries.from_dataframe : Create f

Pipeline to combine multiple data transformers, chaining them together. Parameters ---------- transformers

Parameters ---------- n The width of the window, must be equal to the length of series1 m The

Parameters ---------- window_size The maximum allowed shift between the two series used in DTW.

Abstract class for invertible transformers. All the deriving classes have to implement the static methods :func:`ts_transform()` and

Mixed-data sampling transformer. A transformer that converts higher frequency time series to lower frequency using mixed-data sampling; see

Box-Cox data transformer. See [1]_ for more information about Box-Cox transforms. The transformation is applied independently for ea

Base class for fittable transformers. All the deriving classes have to implement the static methods :func:`ts_transform()` and :func:

Abstract class for data transformers. All the deriving classes have to implement the static method :func:`ts_transform`; this implemented

Data transformer to apply a custom function and its inverse to a (sequence of) ``TimeSeries`` (similar to calling :func:`TimeSeries.m

A transformer that applies window transformation to a TimeSeries or a Sequence of TimeSeries. It expects a dictionary or a list of di

r"""Differencing data transformer. Differencing is typically applied to a time series to make it stationary; see [1]_ for further details.

Generic wrapper class for scalers/encoders/transformers of static covariates. This transformer acts only on static covariates of the series pa

Data transformer to fill missing values from a (sequence of) deterministic ``TimeSeries``. Parameters ---------- fill

Generic wrapper class for using scalers on time series. The underlying `scaler` has to implement the ``fit()``, ``transform()`` and `

MinT Reconcilator. This implements the MinT reconciliation approach presented in [1]_ and summarised in [2]_. Param

Cyclic index encoding for `TimeSeries` that have a time index of type `pd.DatetimeIndex`. Parameters ---------- inde

Parameters ---------- attribute The attribute of the underlying pd.DatetimeIndex from for which to apply cyclic

Parameters ---------- index_generator An instance of `CovariatesIndexGenerator` with methods `generate_train_idx(

Parameters ---------- attribute The attribute of the underlying pd.DatetimeIndex for which to add scalar informat

Parameters ---------- attribute The attribute of the underlying pd.DatetimeIndex for which to add scalar informat

Parameters ---------- index_generator An instance of `CovariatesIndexGenerator` with methods `generate_train_idx(

Parameters ---------- attribute Currently only 'relative' is supported. The generated encoded values will range f

Parameters ---------- attribute Currently only 'relative' is supported. The generated encoded values will range f

Parameters ---------- index_generator An instance of `CovariatesIndexGenerator` with methods `generate_train_idx(

Parameters ---------- attribute A callable that takes an index `index` of type `(pd.DatetimeIndex, pd.RangeIndex)

Parameters ---------- attribute A callable that takes an index `index` of type `(pd.DatetimeIndex, pd.RangeIndex)

SequentialEncoder automatically creates encoder objects from parameter `add_encoders`. `add_encoders` can also be set directly in all

:class:`CovariatesIndexGenerator` generates a time index for covariates at training and inference / prediction time with methods :func:`genera

Single encoders take an `index_generator` to generate the required index for encoding past and future covariates. See darts.utils.data

Parameters ---------- transformer A `FittableDataTransformer` object with a `fit_transform()` and `transform()` m

sMAPE loss as defined in https://robjhyndman.com/hyndsight/smape/ (Chen and Yang 2004) Given a time series of actual values :math:`y

MAPE loss as defined in: https://en.wikipedia.org/wiki/Mean_absolute_percentage_error. Given a time series of actual values :math:`y

Gaussian distribution [1]_. Parameters ---------- n_outputs The number of predicted outputs per model ca

Poisson distribution [1]_. Parameters ---------- n_outputs The number of predicted outputs per model cal

Quantile Regression [1]_. Parameters ---------- n_outputs The number of predicted outputs per model call

Multi-output Quantile Regression [1]_. This is an extension of :class:`QuantileRegression` for scikit-learn models that natively sup

Class probability likelihood. Likelihood to predict the probability of each class for a forecasting classification task. Par

Quantile Prediction Likelihood Can be used to generate quantile predictions for any Darts model that wraps a `statsforecast` model.

Base class for all likelihoods. * likelihoods for torch models * likelihoods for sklearn-like models (e.g. SKLearnModel subc

Abstract class for torch likelihood models. Parameters ---------- likelihood_type A pre-defined `LikelihoodType`.

Univariate Gaussian distribution. https://en.wikipedia.org/wiki/Normal_distribution Instead of the pure negative log likeli

Poisson distribution. Can typically be used to model event counts during time intervals, when the events happen independently of the

Negative Binomial distribution. https://en.wikipedia.org/wiki/Negative_binomial_distribution It does not support priors.

Bernoulli distribution. https://en.wikipedia.org/wiki/Bernoulli_distribution - Univariate discrete distribution. -

Cauchy Distribution. https://en.wikipedia.org/wiki/Cauchy_distribution - Univariate continuous distribution. - Supp

Continuous Bernoulli distribution. https://en.wikipedia.org/wiki/Continuous_Bernoulli_distribution - Univariate continuous

Dirichlet distribution. https://en.wikipedia.org/wiki/Dirichlet_distribution - Multivariate continuous distribution, modeli

Exponential distribution. https://en.wikipedia.org/wiki/Exponential_distribution - Univariate continuous distribution.

Gamma distribution. https://en.wikipedia.org/wiki/Gamma_distribution - Univariate continuous distribution - Support

Geometric distribution. https://en.wikipedia.org/wiki/Geometric_distribution - Univariate discrete distribution - S

Gumbel distribution. https://en.wikipedia.org/wiki/Gumbel_distribution - Univariate continuous distribution - Suppo

Half-normal distribution. https://en.wikipedia.org/wiki/Half-normal_distribution - Univariate continuous distribution.

Laplace distribution. https://en.wikipedia.org/wiki/Laplace_distribution - Univariate continuous distribution - Sup

Log-normal distribution. https://en.wikipedia.org/wiki/Log-normal_distribution - Univariate continuous distribution.

Weibull distribution. https://en.wikipedia.org/wiki/Weibull_distribution - Univariate continuous distribution - Sup

The "likelihood" corresponding to quantile regression. It uses the Quantile Loss Metric for custom quantiles centered around q=0.5.

Sequential Inference Dataset Each sample drawn from this dataset is an eight-element tuple extracted from a specific time window and

Abstract class for all datasets that can be used with Darts' `TorchForecastingModel`. Provides an efficient method to compute the fe

Shifted Training Dataset Each sample drawn from this dataset is a seven-element tuple extracted from a specific time window and set o

Sequential Training Dataset Each sample drawn from this dataset is a seven-element tuple extracted from a specific time window and se

Horizon Based Training Dataset A dataset inspired by the N-BEATS way of training on the M4 dataset: https://arxiv.org/abs/1905.10437.