def _make_fh(cutoff, steps, fh_type, is_relative):
"""Helper function to construct forecasting horizons for testing"""
from sktime.forecasting.tests._config import INDEX_TYPE_LOOKUP
fh_class = INDEX_TYPE_LOOKUP[fh_type]
if isinstance(steps, (int, np.integer)):
steps = np.array([steps], dtype=np.int)
if is_relative:
return ForecastingHorizon(fh_class(steps), is_relative=is_relative)
else:
kwargs = {}
if fh_type == "datetime":
steps *= cutoff.freq
if fh_type == "period":
kwargs = {"freq": cutoff.freq}
values = cutoff + steps
return ForecastingHorizon(fh_class(values, **kwargs), is_relative)
def _transform(self, X, y=None):
"""Transform X and return a transformed version.
private _transform containing the core logic, called from transform
Parameters
----------
X : pd.Series or pd.DataFrame
Data to be transformed
y : pd.DataFrame, default=None
Additional data, e.g., labels for transformation
Returns
-------
Xt : pd.Series or pd.DataFrame, same type as X
transformed version of X, detrended series
"""
fh = ForecastingHorizon(X.index, is_relative=False)
# univariate: X is pd.Series
if isinstance(X, pd.Series):
# note: the y in the transformer is exogeneous in the forecaster, i.e., X
X_pred = self.forecaster_.predict(fh=fh, X=y)
Xt = X - X_pred
return Xt
# multivariate: X is pd.DataFrame
elif isinstance(X, pd.DataFrame):
Xt = X.copy()
# check if all columns are known
X_fit_keys = set(self.forecaster_.keys())
X_new_keys = set(X.columns)
difference = X_new_keys.difference(X_fit_keys)
if len(difference) != 0:
raise ValueError("X contains columns that have not been "
"seen in fit: " + str(difference))
for colname in Xt.columns:
X_pred = self.forecaster_[colname].predict(fh=fh, X=y)
Xt[colname] = Xt[colname] - X_pred
return Xt
else:
raise TypeError("X must be pd.Series or pd.DataFrame")
def transform(self, Z, X=None):
"""
Remove trend from the data.
Parameters
----------
y : pd.Series
Time series to be detrended
X : pd.DataFrame, optional (default=False)
Exogenous variables
Returns
-------
y_hat : pd.Series
De-trended series
"""
self.check_is_fitted()
z = check_series(Z, enforce_univariate=True)
fh = ForecastingHorizon(z.index, is_relative=False)
z_pred = self.forecaster_.predict(fh, X)
return z - z_pred
def inverse_transform(self, Z, X=None):
"""
Add trend back to a time series
Parameters
----------
y : pd.Series, list
Detrended time series to revert
X : pd.DataFrame, optional (default=False)
Exogenous variables
Returns
-------
y_hat : pd.Series
Series with the trend
"""
self.check_is_fitted()
z = check_series(Z, enforce_univariate=True)
fh = ForecastingHorizon(z.index, is_relative=False)
z_pred = self.forecaster_.predict(fh, X)
return z + z_pred
def inverse_transform(self, y, X=None):
"""
Add trend back to a time series
Parameters
----------
y : pd.Series, list
Detrended time series to revert
X : pd.DataFrame, optional (default=False)
Exogenous variables
Returns
-------
y_hat : pd.Series
Series with the trend
"""
self.check_is_fitted()
y = check_y(y)
fh = ForecastingHorizon(y.index, is_relative=False)
y_pred = self.forecaster_.predict(fh=fh, X=X)
return y + y_pred
def _transform(self, X, y=None):
"""Transform X and return a transformed version.
private _transform containing the core logic, called from transform
Parameters
----------
X : pd.Series or pd.DataFrame
Data to be transformed
y : pd.DataFrame, default=None
Additional data, e.g., labels for transformation
Returns
-------
Xt : pd.Series or pd.DataFrame, same type as X
transformed version of X, detrended series
"""
z = X
fh = ForecastingHorizon(z.index, is_relative=False)
# multivariate
if isinstance(z, pd.DataFrame):
z = z.copy()
# check if all columns are known
Z_fit_keys = set(self.forecaster_.keys())
Z_new_keys = set(z.columns)
difference = Z_new_keys.difference(Z_fit_keys)
if len(difference) != 0:
raise ValueError("Z contains columns that have not been "
"seen in fit: " + str(difference))
for colname in z.columns:
z_pred = self.forecaster_[colname].predict(fh, y)
z[colname] = z[colname] - z_pred
Xt = z
# univariate
else:
z_pred = self.forecaster_.predict(fh, y)
Xt = z - z_pred
return Xt
def inverse_transform(self, Z, X=None):
"""Add trend back to a time series.
Parameters
----------
y : pd.Series, list
Detrended time series to revert.
X : pd.DataFrame, optional (default=False)
Exogenous variables.
Returns
-------
y_hat : pd.Series
Series with the trend.
"""
self.check_is_fitted()
z = check_series(Z)
fh = ForecastingHorizon(z.index, is_relative=False)
# multivariate
if isinstance(z, pd.DataFrame):
z = z.copy()
# check if all columns are known
Z_fit_keys = set(self.forecaster_.keys())
Z_new_keys = set(z.columns)
difference = Z_new_keys.difference(Z_fit_keys)
if len(difference) != 0:
raise ValueError(
"Z contains columns that have not been "
"seen in fit: " + difference
)
for colname in z.columns:
z_pred = self.forecaster_[colname].predict(fh, X)
z[colname] = z[colname] + z_pred
return z
# univariate
else:
z_pred = self.forecaster_.predict(fh, X)
return z + z_pred
def transform(self, Z, X=None):
"""Remove trend from the data.
Parameters
----------
y : pd.Series
Time series to be detrended.
X : pd.DataFrame, optional (default=False)
Exogenous variables.
Returns
-------
y_hat : pd.Series
De-trended series.
"""
self.check_is_fitted()
z = check_series(Z)
fh = ForecastingHorizon(z.index, is_relative=False)
# multivariate
if isinstance(z, pd.DataFrame):
z = z.copy()
# check if all columns are known
Z_fit_keys = set(self.forecaster_.keys())
Z_new_keys = set(z.columns)
difference = Z_new_keys.difference(Z_fit_keys)
if len(difference) != 0:
raise ValueError(
"Z contains columns that have not been "
"seen in fit: " + str(difference)
)
for colname in z.columns:
z_pred = self.forecaster_[colname].predict(fh, X)
z[colname] = z[colname] - z_pred
return z
# univariate
else:
z_pred = self.forecaster_.predict(fh, X)
return z - z_pred
def transform(self, y, X=None):
"""
Remove trend from the data.
Parameters
----------
y : pd.Series, list
Time series to be detrended
X : pd.DataFrame, optional (default=False)
Exogenous variables
Returns
-------
y_hat : pd.Series
De-trended series
"""
self.check_is_fitted()
y = check_y(y)
fh = ForecastingHorizon(y.index, is_relative=False)
y_pred = self.forecaster_.predict(fh, X=X)
return y - y_pred
def _inverse_transform(self, X, y=None):
"""Logic used by `inverse_transform` to reverse transformation on `X`.
Parameters
----------
X : pd.Series or pd.DataFrame
Data to be inverse transformed
y : pd.DataFrame, default=None
Additional data, e.g., labels for transformation
Returns
-------
Xt : pd.Series or pd.DataFrame, same type as X
inverse transformed version of X
"""
fh = ForecastingHorizon(X.index, is_relative=False)
# univariate: X is pd.Series
if isinstance(X, pd.Series):
# note: the y in the transformer is exogeneous in the forecaster, i.e., X
X_pred = self.forecaster_.predict(fh=fh, X=y)
return X + X_pred
# multivariate: X is pd.DataFrame
if isinstance(X, pd.DataFrame):
X = X.copy()
# check if all columns are known
X_fit_keys = set(self.forecaster_.keys())
X_new_keys = set(X.columns)
difference = X_new_keys.difference(X_fit_keys)
if len(difference) != 0:
raise ValueError("X contains columns that have not been "
"seen in fit: " + str(difference))
for colname in X.columns:
X_pred = self.forecaster_[colname].predict(fh=fh, X=y)
X[colname] = X[colname] + X_pred
return X
def _inverse_transform(self, X, y=None):
"""Logic used by `inverse_transform` to reverse transformation on `X`.
Parameters
----------
X : pd.Series or pd.DataFrame
Data to be inverse transformed
y : pd.DataFrame, default=None
Additional data, e.g., labels for transformation
Returns
-------
Xt : pd.Series or pd.DataFrame, same type as X
inverse transformed version of X
"""
z = X
fh = ForecastingHorizon(z.index, is_relative=False)
# multivariate
if isinstance(z, pd.DataFrame):
z = z.copy()
# check if all columns are known
Z_fit_keys = set(self.forecaster_.keys())
Z_new_keys = set(z.columns)
difference = Z_new_keys.difference(Z_fit_keys)
if len(difference) != 0:
raise ValueError("Z contains columns that have not been "
"seen in fit: " + difference)
for colname in z.columns:
z_pred = self.forecaster_[colname].predict(fh, X)
z[colname] = z[colname] + z_pred
return z
# univariate
else:
z_pred = self.forecaster_.predict(fh, X)
return z + z_pred