Skip to content

preprocessing

TimeSeriesDifferentiator(order=1)

Bases: BaseEstimator, TransformerMixin

Transforms a time series into a differentiated time series of order n. It also reverts the differentiation.

Parameters:

Name Type Description Default
order int

Order of differentiation.

1

Attributes:

Name Type Description
order int

Order of differentiation.

initial_values list

List with the initial value of the time series after each differentiation. This is used to revert the differentiation.

last_values list

List with the last value of the time series after each differentiation. This is used to revert the differentiation of a new window of data. A new window of data is a time series that starts right after the time series used to fit the transformer.

Source code in skforecast\preprocessing\preprocessing.py
78
79
80
81
82
83
84
85
86
87
88
89
90
def __init__(
    self, 
    order: int = 1
) -> None:

    if not isinstance(order, int):
        raise TypeError(f"Parameter 'order' must be an integer greater than 0. Found {type(order)}.")
    if order < 1:
        raise ValueError(f"Parameter 'order' must be an integer greater than 0. Found {order}.")

    self.order = order
    self.initial_values = []
    self.last_values = []

fit(X, y=None)

Fits the transformer. This method only removes the values stored in self.initial_values.

Parameters:

Name Type Description Default
X numpy ndarray

Time series to be differentiated.

required
y Ignored

Not used, present here for API consistency by convention.

None

Returns:

Name Type Description
self TimeSeriesDifferentiator
Source code in skforecast\preprocessing\preprocessing.py
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
@_check_X_numpy_ndarray_1d
def fit(
    self, 
    X: np.ndarray, 
    y: Any = None
) -> Self:
    """
    Fits the transformer. This method only removes the values stored in
    `self.initial_values`.

    Parameters
    ----------
    X : numpy ndarray
        Time series to be differentiated.
    y : Ignored
        Not used, present here for API consistency by convention.

    Returns
    -------
    self : TimeSeriesDifferentiator

    """

    self.initial_values = []
    self.last_values = []

    for i in range(self.order):
        if i == 0:
            self.initial_values.append(X[0])
            self.last_values.append(X[-1])
            X_diff = np.diff(X, n=1)
        else:
            self.initial_values.append(X_diff[0])
            self.last_values.append(X_diff[-1])
            X_diff = np.diff(X_diff, n=1)

    return self

transform(X, y=None)

Transforms a time series into a differentiated time series of order n and stores the values needed to revert the differentiation.

Parameters:

Name Type Description Default
X numpy ndarray

Time series to be differentiated.

required
y Ignored

Not used, present here for API consistency by convention.

None

Returns:

Name Type Description
X_diff numpy ndarray

Differentiated time series. The length of the array is the same as the original time series but the first n=order values are nan.

Source code in skforecast\preprocessing\preprocessing.py
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
@_check_X_numpy_ndarray_1d
def transform(
    self, 
    X: np.ndarray, 
    y: Any = None
) -> np.ndarray:
    """
    Transforms a time series into a differentiated time series of order n and
    stores the values needed to revert the differentiation.

    Parameters
    ----------
    X : numpy ndarray
        Time series to be differentiated.
    y : Ignored
        Not used, present here for API consistency by convention.

    Returns
    -------
    X_diff : numpy ndarray
        Differentiated time series. The length of the array is the same as
        the original time series but the first n=`order` values are nan.

    """

    X_diff = np.diff(X, n=self.order)
    X_diff = np.append((np.full(shape=self.order, fill_value=np.nan)), X_diff)

    return X_diff

inverse_transform(X, y=None)

Reverts the differentiation. To do so, the input array is assumed to be a differentiated time series of order n that starts right after the the time series used to fit the transformer.

Parameters:

Name Type Description Default
X numpy ndarray

Differentiated time series.

required
y Ignored

Not used, present here for API consistency by convention.

None

Returns:

Name Type Description
X_diff numpy ndarray

Reverted differentiated time series.

Source code in skforecast\preprocessing\preprocessing.py
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
@_check_X_numpy_ndarray_1d
def inverse_transform(
    self, 
    X: np.ndarray, 
    y: Any = None
) -> np.ndarray:
    """
    Reverts the differentiation. To do so, the input array is assumed to be
    a differentiated time series of order n that starts right after the
    the time series used to fit the transformer.

    Parameters
    ----------
    X : numpy ndarray
        Differentiated time series.
    y : Ignored
        Not used, present here for API consistency by convention.

    Returns
    -------
    X_diff : numpy ndarray
        Reverted differentiated time series.

    """

    # Remove initial nan values if present
    X = X[np.argmax(~np.isnan(X)):]
    for i in range(self.order):
        if i == 0:
            X_undiff = np.insert(X, 0, self.initial_values[-1])
            X_undiff = np.cumsum(X_undiff, dtype=float)
        else:
            X_undiff = np.insert(X_undiff, 0, self.initial_values[-(i + 1)])
            X_undiff = np.cumsum(X_undiff, dtype=float)

    return X_undiff

inverse_transform_next_window(X, y=None)

Reverts the differentiation. The input array x is assumed to be a differentiated time series of order n that starts right after the the time series used to fit the transformer.

Parameters:

Name Type Description Default
X numpy ndarray

Differentiated time series. It is assumed o start right after the time series used to fit the transformer.

required
y Ignored

Not used, present here for API consistency by convention.

None

Returns:

Name Type Description
X_undiff numpy ndarray

Reverted differentiated time series.

Source code in skforecast\preprocessing\preprocessing.py
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
@_check_X_numpy_ndarray_1d
def inverse_transform_next_window(
    self,
    X: np.ndarray,
    y: Any = None
) -> np.ndarray:
    """
    Reverts the differentiation. The input array `x` is assumed to be a 
    differentiated time series of order n that starts right after the
    the time series used to fit the transformer.

    Parameters
    ----------
    X : numpy ndarray
        Differentiated time series. It is assumed o start right after
        the time series used to fit the transformer.
    y : Ignored
        Not used, present here for API consistency by convention.

    Returns
    -------
    X_undiff : numpy ndarray
        Reverted differentiated time series.

    """

    # Remove initial nan values if present
    X = X[np.argmax(~np.isnan(X)):]

    for i in range(self.order):
        if i == 0:
            X_undiff = np.cumsum(X, dtype=float) + self.last_values[-1]
        else:
            X_undiff = np.cumsum(X_undiff, dtype=float) + self.last_values[-(i + 1)]

    return X_undiff

series_long_to_dict(data, series_id, index, values, freq)

Convert long format series to dictionary.

Parameters:

Name Type Description Default
data DataFrame

Long format series.

required
series_id str

Column name with the series identifier.

required
index str

Column name with the time index.

required
values str

Column name with the values.

required
freq str

Frequency of the series.

required

Returns:

Name Type Description
series_dict dict

Dictionary with the series.

Source code in skforecast\preprocessing\preprocessing.py
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
def series_long_to_dict(
    data: pd.DataFrame,
    series_id: str,
    index: str,
    values: str,
    freq: str,
) -> dict:
    """
    Convert long format series to dictionary.

    Parameters
    ----------
    data: pandas DataFrame
        Long format series.
    series_id: str
        Column name with the series identifier.
    index: str
        Column name with the time index.
    values: str
        Column name with the values.
    freq: str
        Frequency of the series.

    Returns
    -------
    series_dict: dict
        Dictionary with the series.

    """

    if not isinstance(data, pd.DataFrame):
        raise TypeError("`data` must be a pandas DataFrame.")

    for col in [series_id, index, values]:
        if col not in data.columns:
            raise ValueError(f"Column '{col}' not found in `data`.")

    series_dict = {}
    for k, v in data.groupby(series_id):
        series_dict[k] = v.set_index(index)[values].asfreq(freq).rename(k)
        series_dict[k].index.name = None

    return series_dict

exog_long_to_dict(data, series_id, index, freq, dropna=False)

Convert long format exogenous variables to dictionary.

Parameters:

Name Type Description Default
data DataFrame

Long format exogenous variables.

required
series_id str

Column name with the series identifier.

required
index str

Column name with the time index.

required
freq str

Frequency of the series.

required
dropna bool

If True, drop columns with all values as NaN. This is useful when there are series without some exogenous variables.

False

Returns:

Name Type Description
exog_dict dict

Dictionary with the exogenous variables.

Source code in skforecast\preprocessing\preprocessing.py
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
def exog_long_to_dict(
    data: pd.DataFrame,
    series_id: str,
    index: str,
    freq: str,
    dropna: bool = False,
) -> dict:
    """
    Convert long format exogenous variables to dictionary.

    Parameters
    ----------
    data: pandas DataFrame
        Long format exogenous variables.
    series_id: str
        Column name with the series identifier.
    index: str
        Column name with the time index.
    freq: str
        Frequency of the series.
    dropna: bool, default `False`
        If True, drop columns with all values as NaN. This is useful when
        there are series without some exogenous variables.

    Returns
    -------
    exog_dict: dict
        Dictionary with the exogenous variables.

    """

    if not isinstance(data, pd.DataFrame):
        raise TypeError("`data` must be a pandas DataFrame.")

    for col in [series_id, index]:
        if col not in data.columns:
            raise ValueError(f"Column '{col}' not found in `data`.")

    exog_dict = dict(tuple(data.groupby(series_id)))
    exog_dict = {
        k: v.set_index(index).asfreq(freq).drop(columns=series_id)
        for k, v in exog_dict.items()
    }

    for k in exog_dict.keys():
        exog_dict[k].index.name = None

    if dropna:
        exog_dict = {k: v.dropna(how="all", axis=1) for k, v in exog_dict.items()}

    return exog_dict

DateTimeFeatureTransformer(features=None, encoding='cyclical', max_values=None)

Bases: BaseEstimator, TransformerMixin

A transformer for extracting datetime features from the DateTime index of a pandas DataFrame or Series. It can also apply encoding to the extracted features.

Parameters:

Name Type Description Default
features list

List of calendar features (strings) to extract from the index. When None, the following features are extracted: 'year', 'month', 'week', 'day_of_week', 'day_of_month', 'day_of_year', 'weekend', 'hour', 'minute', 'second'.

`None`
encoding str

Encoding method for the extracted features. Options are None, 'cyclical' or 'onehot'.

`'cyclical'`
max_values dict

Dictionary of maximum values for the cyclical encoding of calendar features. When None, the following values are used: {'month': 12, 'week': 52, 'day_of_week': 7, 'day_of_month': 31, 'day_of_year': 365, 'hour': 24, 'minute': 60, 'second': 60}.

`None`

Attributes:

Name Type Description
features list

List of calendar features to extract from the index.

encoding str

Encoding method for the extracted features.

max_values dict

Dictionary of maximum values for the cyclical encoding of calendar features.

Source code in skforecast\preprocessing\preprocessing.py
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
def __init__(
    self,
    features: Optional[list] = None,
    encoding: str = "cyclical",
    max_values: Optional[dict] = None
) -> None:

    if encoding not in ["cyclical", "onehot", None]:
        raise ValueError("Encoding must be one of 'cyclical', 'onehot' or None")

    self.features = (
        features
        if features is not None
        else [
            "year",
            "month",
            "week",
            "day_of_week",
            "day_of_month",
            "day_of_year",
            "weekend",
            "hour",
            "minute",
            "second",
        ]
    )
    self.encoding = encoding
    self.max_values = (
        max_values
        if max_values is not None
        else {
            "month": 12,
            "week": 52,
            "day_of_week": 7,
            "day_of_month": 31,
            "day_of_year": 365,
            "hour": 24,
            "minute": 60,
            "second": 60,
        }
    )

fit(X, y=None)

A no-op method to satisfy the scikit-learn API.

Source code in skforecast\preprocessing\preprocessing.py
520
521
522
523
524
def fit(self, X, y=None):
    """
    A no-op method to satisfy the scikit-learn API.
    """
    return self

transform(X)

Create datetime features from the DateTime index of a pandas DataFrame or Series.

Parameters:

Name Type Description Default
X pandas Series, pandas DataFrame

Input DataFrame or Series with a datetime index.

required

Returns:

Name Type Description
X_new pandas DataFrame

DataFrame with the extracted (and optionally encoded) datetime features.

Source code in skforecast\preprocessing\preprocessing.py
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
def transform(
    self,
    X: Union[pd.Series, pd.DataFrame]
) -> pd.DataFrame:
    """
    Create datetime features from the DateTime index of a pandas DataFrame or Series.

    Parameters
    ----------
    X : pandas Series, pandas DataFrame
        Input DataFrame or Series with a datetime index.

    Returns
    -------
    X_new : pandas DataFrame
        DataFrame with the extracted (and optionally encoded) datetime features.

    """

    X_new = create_datetime_features(
                X          = X,
                encoding   = self.encoding,
                features   = self.features,
                max_values = self.max_values,
            )

    return X_new

create_datetime_features(X, features=None, encoding='cyclical', max_values=None)

Extract datetime features from the DateTime index of a pandas DataFrame or Series.

Parameters:

Name Type Description Default
X pandas Series, pandas DataFrame

Input DataFrame or Series with a datetime index.

required
features list

List of calendar features (strings) to extract from the index. When None, the following features are extracted: 'year', 'month', 'week', 'day_of_week', 'day_of_month', 'day_of_year', 'weekend', 'hour', 'minute', 'second'.

`None`
encoding str

Encoding method for the extracted features. Options are None, 'cyclical' or 'onehot'.

`'cyclical'`
max_values dict

Dictionary of maximum values for the cyclical encoding of calendar features. When None, the following values are used: {'month': 12, 'week': 52, 'day_of_week': 7, 'day_of_month': 31, 'day_of_year': 365, 'hour': 24, 'minute': 60, 'second': 60}.

`None`

Returns:

Name Type Description
X_new pandas DataFrame

DataFrame with the extracted (and optionally encoded) datetime features.

Source code in skforecast\preprocessing\preprocessing.py
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
def create_datetime_features(
    X: Union[pd.Series, pd.DataFrame],
    features: Optional[list] = None,
    encoding: str = "cyclical",
    max_values: Optional[dict] = None,
) -> pd.DataFrame:
    """
    Extract datetime features from the DateTime index of a pandas DataFrame or Series.

    Parameters
    ----------
    X : pandas Series, pandas DataFrame
        Input DataFrame or Series with a datetime index.
    features : list, default `None`
        List of calendar features (strings) to extract from the index. When `None`,
        the following features are extracted: 'year', 'month', 'week', 'day_of_week',
        'day_of_month', 'day_of_year', 'weekend', 'hour', 'minute', 'second'.
    encoding : str, default `'cyclical'`
        Encoding method for the extracted features. Options are None, 'cyclical' or
        'onehot'.
    max_values : dict, default `None`
        Dictionary of maximum values for the cyclical encoding of calendar features.
        When `None`, the following values are used: {'month': 12, 'week': 52, 
        'day_of_week': 7, 'day_of_month': 31, 'day_of_year': 365, 'hour': 24, 
        'minute': 60, 'second': 60}.

    Returns
    -------
    X_new : pandas DataFrame
        DataFrame with the extracted (and optionally encoded) datetime features.

    """

    if not isinstance(X, (pd.DataFrame, pd.Series)):
        raise TypeError("Input `X` must be a pandas Series or DataFrame")
    if not isinstance(X.index, pd.DatetimeIndex):
        raise TypeError("Input `X` must have a pandas DatetimeIndex")
    if encoding not in ["cyclical", "onehot", None]:
        raise ValueError("Encoding must be one of 'cyclical', 'onehot' or None")

    default_features = [
        "year",
        "month",
        "week",
        "day_of_week",
        "day_of_month",
        "day_of_year",
        "weekend",
        "hour",
        "minute",
        "second",
    ]
    features = features or default_features

    default_max_values = {
        "month": 12,
        "week": 52,
        "day_of_week": 7,
        "day_of_month": 31,
        "day_of_year": 365,
        "hour": 24,
        "minute": 60,
        "second": 60,
    }
    max_values = max_values or default_max_values

    X_new = pd.DataFrame(index=X.index)

    datetime_attrs = {
        "year": "year",
        "month": "month",
        "week": lambda idx: idx.isocalendar().week,
        "day_of_week": "dayofweek",
        "day_of_year": "dayofyear",
        "day_of_month": "day",
        "weekend": lambda idx: (idx.weekday >= 5).astype(int),
        "hour": "hour",
        "minute": "minute",
        "second": "second",
    }

    not_supported_features = set(features) - set(datetime_attrs.keys())
    if not_supported_features:
        raise ValueError(
            f"Features {not_supported_features} are not supported. "
            f"Supported features are {list(datetime_attrs.keys())}."
        )

    for feature in features:
        attr = datetime_attrs[feature]
        X_new[feature] = (
            attr(X.index) if callable(attr) else getattr(X.index, attr).astype(int)
        )

    if encoding == "cyclical":
        cols_to_drop = []
        for feature, max_val in max_values.items():
            if feature in X_new.columns:
                X_new[f"{feature}_sin"] = np.sin(2 * np.pi * X_new[feature] / max_val)
                X_new[f"{feature}_cos"] = np.cos(2 * np.pi * X_new[feature] / max_val)
                cols_to_drop.append(feature)
        X_new = X_new.drop(columns=cols_to_drop)
    elif encoding == "onehot":
        X_new = pd.get_dummies(
            X_new, columns=features, drop_first=False, sparse=False, dtype=int
        )

    return X_new