Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

[DOC] added post-processing notes in tutorial #3878

Merged
merged 9 commits into from Dec 20, 2022
Merged

[DOC] added post-processing notes in tutorial #3878

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
9 commits
Select commit Hold shift + click to select a range
b157019
added additional notes on postprocessing usage after estimation
nshahpazov Dec 3, 2022
670490c
added nshahpazov to contributors
nshahpazov Dec 3, 2022
222a555
reverted back, but keeping only necessary changes
nshahpazov Dec 3, 2022
60a591c
added output of the new example
nshahpazov Dec 3, 2022
a42a6d4
rephrased explanation of different types of pipelines (pre and post p…
nshahpazov Dec 4, 2022
1d10740
removed trailing space
nshahpazov Dec 4, 2022
e647ef7
fixed a typo
nshahpazov Dec 5, 2022
9ddddee
minor sharpening
fkiraly Dec 19, 2022
d96bfb4
revert metadata
fkiraly Dec 19, 2022
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
71 changes: 39 additions & 32 deletions .all-contributorsrc
View file
Open in desktop
Expand Up @@ -1652,7 +1652,6 @@
]
},
{

"login": "AurumnPegasus",
"name": "Shivansh Subramanian",
"avatar_url": "https://avatars.githubusercontent.com/u/54315149?v=4",
Expand Down Expand Up @@ -1702,7 +1701,6 @@
"profile": "https://github.com/mariamjabara",
"contributions": [
"code"

]
},
{
Expand Down Expand Up @@ -1780,7 +1778,7 @@
]
},
{
"login": "bugslayer-332" ,
"login": "bugslayer-332",
"name": "Arepalli Yashwanth Reddy",
"profile": "https://github.com/bugslayer-332",
"contributions": [
Expand Down Expand Up @@ -1845,7 +1843,7 @@
"code",
"test",
"ideas"
]
]
},
{
"login": "KishManani",
Expand All @@ -1858,31 +1856,40 @@
"test",
"bug",
"ideas"
},
{
"login": "jorenham",
"name": "Joren Hammudoglu",
"profile": "https://github.com/jorenham",
"contributions": [
"build"
]
},
{
"login": "wolph",
"name": "Rick van Hattem",
"profile": "https://github.com/wolph",
"contributions": [
"build"
]
},
{
"login": "templierw",
"name": "William Templier",
"avatar_url": "https://github.com/templierw.png",
"profile": "https://www.linkedin.com/in/templierw/",
"contributions": [
"doc"
]
}
]
}
},
{
"login": "jorenham",
"name": "Joren Hammudoglu",
"profile": "https://github.com/jorenham",
"contributions": [
"build"
]
},
{
"login": "wolph",
"name": "Rick van Hattem",
"profile": "https://github.com/wolph",
"contributions": [
"build"
]
},
{
"login": "templierw",
"name": "William Templier",
"avatar_url": "https://github.com/templierw.png",
"profile": "https://www.linkedin.com/in/templierw/",
"contributions": [
"doc"
]
},
{
"login": "nshahpazov",
"name": "Nikola Shahpazov",
"avatar_url": "https://avatars.githubusercontent.com/nshahpazov",
"profile": "https://www.linkedin.com/in/nshahpazov/",
"contributions": [
"doc"
]
}
]
}
81 changes: 77 additions & 4 deletions examples/01_forecasting.ipynb
View file
Open in desktop
Expand Up @@ -5365,7 +5365,8 @@
"source": [
"#### 3.2.1 The basic forecasting pipeline<a class=\"anchor\" id=\"section_3_2_1\"></a>\n",
"\n",
"`sktime` provides a generic pipeline object for this kind of composite modelling, the `TransforemdTargetForecaster`. It chains an arbitrary number of transformations with a forecaster. The transformations should be instances of estimators with series-to-series-transformer scitype. An example of the syntax is below:"
"`sktime` provides a generic pipeline object for this kind of composite modelling, the `TransforemdTargetForecaster`. It chains an arbitrary number of transformations with a forecaster. The transformations can either be pre-processing transformations or a post-processing transformations. An example of a forecaster with pre-processing\n",
"transformations can be seen below."
]
},
{
Expand Down Expand Up @@ -5425,7 +5426,7 @@
"cell_type": "markdown",
"metadata": {},
"source": [
"The `TransformedTargetForecaster` is constructed with a list of steps, each a pair of name and estimator. The last estimator should be of forecaster scitype, the other estimators should be series-to-series transformers which possess both a `transform` and `inverse_transform` method. The resulting estimator is of forecaster scitype and has all interface defining methods. In `fit`, all transformers apply `fit_transforms` to the data, then the forecaster's `fit`; in `predict`, first the forecaster's `predict` is applied, then the transformers' `inverse_transform` in reverse order."
"In the above example, the `TransformedTargetForecaster` is constructed with a list of steps, each a pair of name and estimator, where the last estimator is a forecaster scitype. The pre-processing transformers should be series-to-series transformers which possess both a `transform` and an `inverse_transform` method. The resulting estimator is of forecaster scitype and has all interface defining methods. In `fit`, all transformers apply `fit_transforms` to the data, then the forecaster's `fit`; in `predict`, first the forecaster's `predict` is applied, then the transformers' `inverse_transform` in reverse order."
nshahpazov marked this conversation as resolved.
Show resolved Hide resolved
]
},
{
Expand Down Expand Up @@ -5471,7 +5472,7 @@
},
{
"cell_type": "code",
"execution_count": 252,
"execution_count": 182,
"metadata": {},
"outputs": [
{
Expand Down Expand Up @@ -5508,7 +5509,7 @@
" 'ARIMA__with_intercept': True}"
]
},
"execution_count": 252,
"execution_count": 182,
"metadata": {},
"output_type": "execute_result"
}
Expand All @@ -5523,6 +5524,78 @@
"forecaster.get_params()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"We can also create a pipeline with post-processing transformations, these are transformations *after* the forecaster, in a dunder pipeline or a `TransformedTargetForecaster`.\n",
"\n",
"Below is an example of a multiple seasonality model, with integer rounding post-processing of the predictions:"
]
},
{
"cell_type": "code",
"execution_count": 186,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1958-01 334.0\n",
"1958-02 338.0\n",
"1958-03 317.0\n",
"Freq: M, dtype: float64"
]
},
"execution_count": 186,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from sktime.transformations.series.func_transform import FunctionTransformer\n",
"\n",
"forecaster = ARIMA() * FunctionTransformer(lambda y: y.round())\n",
"forecaster.fit_predict(y, fh=fh).head(3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Both pre- and post-processing transformers can be present, in this case the post-processing transformations will be applied after the `inverse-transform` of the pre-processing ones."
]
},
{
"cell_type": "code",
"execution_count": 188,
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"1958-01 339.0\n",
"1958-02 334.0\n",
"1958-03 381.0\n",
"Freq: M, dtype: float64"
]
},
"execution_count": 188,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"forecaster = (\n",
" Deseasonalizer(model=\"multiplicative\", sp=12)\n",
" * Deseasonalizer(model=\"multiplicative\", sp=3)\n",
" * ARIMA()\n",
" * FunctionTransformer(lambda y: y.round())\n",
")\n",
"\n",
"forecaster.fit_predict(y_train, fh=fh).head(3)"
]
},
{
"cell_type": "markdown",
"metadata": {},
Expand Down