[BUG] `MSTL.inverse_transform` fails if `return_components=False`
eangius opened this issue · comments
Describe the bug
Sktime is a great library thanks. Perhaps this is user error, but running the MSTL to remove multiple seasonalities from endogenous variables as a standalone component or within a regular pipeline works as expected but fails to generate predictions when within a TransformedTargetForecaster.
To Reproduce
from sktime.forecasting.naive import *
from sktime.forecasting.compose import *
from sktime.transformations.series.detrend import *
from sktime.datasets import load_airline
pipe = ForecastingPipeline(steps=[
('y', TransformedTargetForecaster(steps=[
('decompose', MSTL(periods=12, return_components=False)),
('forecaster', NaiveForecaster()),
]))
]).fit(y=load_airline(), fh=[1, 2, 3])
y_pred = pipe.predict() # exception hereExpected behavior
No exception at predict time. Seems like something is returning a pd.Series & something downstream is expecting a pd.DataFrame
Additional context
Traceback (most recent call last):
File ".venv/lib/python3.9/site-packages/pandas/core/generic.py", line 575, in _get_axis_number
return cls._AXIS_TO_AXIS_NUMBER[axis]
KeyError: 1
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/Applications/PyCharm CE.app/Contents/plugins/python-ce/helpers/pydev/pydevconsole.py", line 364, in runcode
coro = func()
File "<input>", line 12, in <module>
File ".venv/lib/python3.9/site-packages/sktime/forecasting/base/_base.py", line 448, in predict
y_pred = self._predict(fh=fh, X=X_inner)
File “.venv/lib/python3.9/site-packages/sktime/forecasting/compose/_pipeline.py", line 531, in _predict
return self.forecaster_.predict(fh, X)
File ".venv/lib/python3.9/site-packages/sktime/forecasting/base/_base.py", line 448, in predict
y_pred = self._predict(fh=fh, X=X_inner)
File “.venv/lib/python3.9/site-packages/sktime/forecasting/compose/_pipeline.py", line 1053, in _predict
y_pred = self._get_inverse_transform(self.transformers_pre_, y_pred, X)
File ".venv/lib/python3.9/site-packages/sktime/forecasting/compose/_pipeline.py", line 149, in _get_inverse_transform
y = transformer.inverse_transform(y, X)
File ".venv/lib/python3.9/site-packages/sktime/transformations/base.py", line 738, in inverse_transform
Xt = self._inverse_transform(X=X_inner, y=y_inner)
File ".venv/lib/python3.9/site-packages/sktime/transformations/series/detrend/mstl.py", line 203, in _inverse_transform
row_sums = X.sum(axis=1)
File ".venv/lib/python3.9/site-packages/pandas/core/series.py", line 6519, in sum
return NDFrame.sum(self, axis, skipna, numeric_only, min_count, **kwargs)
File ".venv/lib/python3.9/site-packages/pandas/core/generic.py", line 12503, in sum
return self._min_count_stat_function(
File ".venv/lib/python3.9/site-packages/pandas/core/generic.py", line 12486, in _min_count_stat_function
return self._reduce(
File ".venv/lib/python3.9/site-packages/pandas/core/series.py", line 6430, in _reduce
self._get_axis_number(axis)
File ".venv/lib/python3.9/site-packages/pandas/core/generic.py", line 577, in _get_axis_number
raise ValueError(f"No axis named {axis} for object type {cls.__name__}")
ValueError: No axis named 1 for object type SeriesVersions
How odd.
Strictly speaking it is superfluous to wrap the TransformedTargetForecaster in ForecatsingPipeline, but that should not impact behaviour (it just should be ignored, because it is a single elemen tpipeline).
I can see the issue: it is the axis=1 argument. If return_components=True, then the result will be a pd.DataFrame because it is multivariate. If it is False, then it is a pd.Series.
From a methodological standpoint, the inverse works correctly only in the return_components=True case, if the seasonal component is also forecast.
I see that one would expect that the seasonal component is continued periodically and added back.
So, returning X if pd.Series would remove the exception, but lead to unexpected behaviour, as the seasonal components are not added back.
Perchance, do you know, @eangius, is there an easy way to get an extrapolated form of all seasonal components in statsmodels MSTL? There should be?
Also FYI @luca-miniati who is the author and maintainer.
To clarify, I think inverse_transform should do:
- if
return_components=True, exactly what it does currently - in this case it needs forecasters for all seasonal components and the residual if used in a pipeline - if
return_components=False, a naive periodic continuation should be made for seasonal components, and it should be added to the transformed values. This might be slightly challenging, given that the index seen in_inverse_transformneed not be contiguous with, or could intersect with, the index seen infit.
updated the issue title - imo the root cause is that MSTL.inverse_transform fails whenever return_components=False
there is also a wider issue, namely that inverse_transform test coverage seems insufficient to detect this, which should be investigated.
Hi Franz, long time no see! I'd like to implement the functionality for return_components=False.
Let me know if I understand the solution correctly:
- make predictions of the seasonal time series, using the provided fh
- add up all the values, and return as
pd.Series
And a clarifying question: why would the index seen in fit potentially not match the index of _inverse_transform?
Thanks for the quick diagnostic @fkiraly. Unfortunately I’m still a knob at statsmodel to tell how to extract all seasonal components..
For context, we are wrapping MSTL into a TransformedTargetForecaster because we have a previous processing step for the exogenous variables but that was not relevant to reproduce the problem.
As an extra bit of context, we tried with return_components=True and filtering out the other columns in a FunctionTransformer to keep things univariate but that gave us a different type of exception..
@eangius, as possible workarounds for de/re-trending in a pipeline:
- you can pipeline multiple
Deseasonalizer-s, likeDeseasonalizer(sp=24) * Deseasonalizer(sp=24*7) * my_forecasterfor daily and weekly (if your data is hourly - you can try
StatsforecastMSTL, this is a forecaster that is optimized and with integrated MSTL, though with a heavier dependency footprint
Hi Franz, long time no see!
Nice to hear from you again, as well!
I'd like to implement the functionality for
return_components=False.
Great, let me know if I can help.
Let me know if I understand the solution correctly:
make predictions of the seasonal time series, using the provided fh
add up all the values, and return as
pd.Series
Yes, this should happen when it is pipelined with a forecaster.
Though, the MSTL estimator is a transformer, so the transformer needs to carry out the transformation steps only.
So, we need to take the indices in _inverse_transform, and determine the periodic pattern implied by what was fitted on fit.
And a clarifying question: why would the index seen in fit potentially not match the index of
_inverse_transform?
If you work out what happens in a forecasting pipeline, the transformer gets the historic indices in fit, e.g., 0, 1, 2, ..., 100, and the indices corresponding to the fh in predict, fore a fh of 1, 2, 3, the X in _inverse_transform would have index 101, 102, 103.
If we have patterns of periodicities 3, 5, 7, denoting the indices of the periodic patterm by 3-0, 3-1, 3-2; 5-0, 5-1, ..., 5-4; 7-0, ..., 7-6, (dashes just for notation, not "minus") then for incides 101, 102, 103 we should forecast, for components, the indices 3-2, 3-0, 3-1; 5-1, 5-2, 5-3; 7-4, 7-5, 7-6.
(in python, we start counting with 0, so X-1 maps onto any index divisible without remainder by X)
I think this already must be done somewhere in transform if return_components=True?