"""
Box-Cox Transformer
-------------------
"""
from typing import Any, Mapping, Optional, Sequence, Union
try:
from typing import Literal
except ImportError:
from typing_extensions import Literal
import numpy as np
import pandas as pd
from scipy.special import inv_boxcox
from scipy.stats import boxcox, boxcox_normmax
from darts.dataprocessing.transformers.fittable_data_transformer import (
FittableDataTransformer,
)
from darts.dataprocessing.transformers.invertible_data_transformer import (
InvertibleDataTransformer,
)
from darts.logging import get_logger, raise_if
from darts.timeseries import TimeSeries
logger = get_logger(__name__)
[docs]class BoxCox(FittableDataTransformer, InvertibleDataTransformer):
def __init__(
self,
name: str = "BoxCox",
lmbda: Optional[
Union[float, Sequence[float], Sequence[Sequence[float]]]
] = None,
optim_method: Literal["mle", "pearsonr"] = "mle",
global_fit: bool = False,
n_jobs: int = 1,
verbose: bool = False,
):
"""Box-Cox data transformer.
See [1]_ for more information about Box-Cox transforms.
The transformation is applied independently for each dimension (component) of the time series.
For stochastic series, it is done jointly over all samples, effectively merging all samples of
a component in order to compute the transform.
Notes
-----
The scaler will not scale the series' static covariates. This has to be done either before constructing the
series, or later on by extracting the covariates, transforming the values and then reapplying them to the
series. For this, see TimeSeries properties `TimeSeries.static_covariates` and method
`TimeSeries.with_static_covariates()`
Parameters
----------
name
A specific name for the transformer
lmbda
The parameter :math:`\\lambda` of the Box-Cox transform. If a single float is given, the same
:math:`\\lambda` value will be used for all dimensions of the series, for all the series.
If a sequence is given, there is one value per component in the series. If a sequence of sequence
is given, there is one value per component for all series.
If `None` given, will automatically find an optimal value of :math:`\\lambda` (for each dimension
of the time series, for each time series) using :func:`scipy.stats.boxcox_normmax`
with ``method=optim_method``.
optim_method
Specifies which method to use to find an optimal value for the lmbda parameter.
Either 'mle' or 'pearsonr'. Ignored if `lmbda` is not `None`.
global_fit
Optionally, whether all `TimeSeries` passed to the `fit()` method should be used to fit
a *single* set of parameters, or if a different set of parameters should be independently fitted
to each provided `TimeSeries`. If `True`, then a `Sequence[TimeSeries]` is passed to `ts_fit`
and a single set of parameters is fitted using all provided `TimeSeries`. If `False`, then
each `TimeSeries` is individually passed to `ts_fit`, and a different set of fitted parameters
if yielded for each of these fitting operations. See `FittableDataTransformer` documentation for
further details.
n_jobs
The number of jobs to run in parallel. Parallel jobs are created only when a ``Sequence[TimeSeries]`` is
passed as input, parallelising operations regarding different ``TimeSeries``. Defaults to `1`
(sequential). Setting the parameter to `-1` means using all the available processors.
Note: for a small amount of data, the parallelisation overhead could end up increasing the total
required amount of time.
verbose
Whether to print operations progress
Examples
--------
>>> from darts.datasets import AirPassengersDataset
>>> from darts.dataprocessing.transformers import BoxCox
>>> series = AirPassengersDataset().load()
>>> transformer = BoxCox(lmbda=0.2)
>>> series_transformed = transformer.fit_transform(series)
>>> print(series_transformed.head())
<TimeSeries (DataArray) (Month: 5, component: 1, sample: 1)>
array([[[7.84735157]],
[[7.98214351]],
[[8.2765364 ]],
[[8.21563229]],
[[8.04749318]]])
Coordinates:
* Month (Month) datetime64[ns] 1949-01-01 1949-02-01 ... 1949-05-01
* component (component) object '#Passengers'
Dimensions without coordinates: sample
References
----------
.. [1] https://otexts.com/fpp2/transformations.html#mathematical-transformations
"""
raise_if(
not isinstance(optim_method, str)
or optim_method not in ["mle", "pearsonr"],
"optim_method parameter must be either 'mle' or 'pearsonr'",
logger,
)
# Define fixed params (i.e. attributes defined before calling `super().__init__`):
self._lmbda = lmbda
self._optim_method = optim_method
if isinstance(lmbda, Sequence) and isinstance(lmbda[0], Sequence):
parallel_params = ("_lmbda",)
else:
parallel_params = False
super().__init__(
name=name,
n_jobs=n_jobs,
verbose=verbose,
parallel_params=parallel_params,
mask_components=True,
global_fit=global_fit,
)
[docs] @staticmethod
def ts_fit(
series: Union[TimeSeries, Sequence[TimeSeries]],
params: Mapping[str, Any],
*args,
**kwargs,
) -> Union[Sequence[float], pd.Series]:
lmbda, method = params["fixed"]["_lmbda"], params["fixed"]["_optim_method"]
# If `global_fit` is `True`, then `series` will be ` Sequence[TimeSeries]`;
# otherwise, `series` is a single `TimeSeries`:
if isinstance(series, TimeSeries):
series = [series]
if lmbda is None:
# Compute optimal lmbda for each dimension of the time series. In this case, the return type is
# an ndarray and not a Sequence
vals = np.concatenate([BoxCox.stack_samples(ts) for ts in series], axis=0)
lmbda = np.apply_along_axis(boxcox_normmax, axis=0, arr=vals, method=method)
elif isinstance(lmbda, Sequence):
raise_if(
len(lmbda) != series[0].width,
"lmbda should have one value per dimension (ie. column or variable) of the time series",
logger,
)
else:
# Replicate lmbda to match dimensions of the time series
lmbda = [lmbda] * series[0].width
return lmbda