У меня есть серия панд, проиндексированная объектами datetime.date и имеющая некоторые значения, которые меня интересуют в моделировании.У меня около 250 строк.Я пытаюсь использовать первые 150 примеров для обучения модели ARIMA с помощью модуля statsmodels (версия 0.9.0), а затем для составления прогноза для 151-го примера (который я также имею истинное значение).Важно отметить, что у меня есть значения только в будние дни, а не в выходные дни.Я надеюсь, что есть хороший способ справиться с этим.
>> close[0:151]
2018-01-02 0.460
2018-01-03 0.465
2018-01-04 0.475
2018-01-05 0.480
2018-01-08 0.475
2018-01-09 0.500
2018-01-10 0.490
2018-01-11 0.505
2018-01-12 0.500
2018-01-15 0.510
2018-01-16 0.535
2018-01-17 0.500
2018-01-18 0.530
2018-01-19 0.540
2018-01-22 0.505
2018-01-23 0.475
2018-01-24 0.440
2018-01-25 0.490
2018-01-29 0.465
2018-01-30 0.465
2018-01-31 0.465
2018-02-01 0.475
2018-02-02 0.475
2018-02-05 0.455
2018-02-06 0.435
2018-02-07 0.455
2018-02-08 0.455
2018-02-09 0.285
2018-02-12 0.290
2018-02-13 0.275
...
2018-06-27 0.250
2018-06-28 0.245
2018-06-29 0.230
2018-07-02 0.220
2018-07-03 0.250
2018-07-04 0.240
2018-07-05 0.235
2018-07-06 0.225
2018-07-09 0.225
2018-07-10 0.220
2018-07-11 0.215
2018-07-12 0.220
2018-07-13 0.210
2018-07-16 0.220
2018-07-17 0.215
2018-07-18 0.235
2018-07-19 0.240
2018-07-20 0.230
2018-07-23 0.235
2018-07-24 0.215
2018-07-25 0.220
2018-07-26 0.215
2018-07-27 0.215
2018-07-30 0.205
2018-07-31 0.200
2018-08-01 0.195
2018-08-02 0.205
2018-08-03 0.215
2018-08-06 0.215
2018-08-07 0.210
Name: Close, Length: 151, dtype: float64
Может ли кто-нибудь помочь мне с использованием первых 150 строк для обучения модели ARIMA и распечатать прогноз для 151-й строки?
Пытаясь следовать тому, что сделали другие люди, япробовал это:
from statsmodels.tsa.arima_model import ARIMA
model = ARIMA(close[:150], order=(1,1,1))
model_fit = model.fit(disp=0)
Эта модель тренируется, но потом, когда я пытаюсь:
pred = model_fit.predict(start=close.index[150], dynamic=False, typ='levels')
В ответ я получаю следующую ошибку:
---------------------------------------------------------------------------
KeyError Traceback (most recent call last)
~/anaconda3/lib/python3.7/site-packages/pandas/core/indexes/base.py in get_loc(self, key, method, tolerance)
2656 try:
-> 2657 return self._engine.get_loc(key)
2658 except KeyError:
pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()
pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()
pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.get_item()
pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.get_item()
KeyError: datetime.date(2018, 8, 7)
During handling of the above exception, another exception occurred:
KeyError Traceback (most recent call last)
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/base/tsa_model.py in _get_index_label_loc(self, key, base_index)
415 if not isinstance(key, (int, long, np.integer)):
--> 416 loc = self.data.row_labels.get_loc(key)
417 else:
~/anaconda3/lib/python3.7/site-packages/pandas/core/indexes/base.py in get_loc(self, key, method, tolerance)
2658 except KeyError:
-> 2659 return self._engine.get_loc(self._maybe_cast_indexer(key))
2660 indexer = self.get_indexer([key], method=method, tolerance=tolerance)
pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()
pandas/_libs/index.pyx in pandas._libs.index.IndexEngine.get_loc()
pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.get_item()
pandas/_libs/hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.get_item()
KeyError: datetime.date(2018, 8, 7)
During handling of the above exception, another exception occurred:
KeyError Traceback (most recent call last)
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/base/tsa_model.py in _get_prediction_index(self, start, end, index, silent)
476 try:
--> 477 start, start_index, start_oos = self._get_index_label_loc(start)
478 except KeyError:
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/base/tsa_model.py in _get_index_label_loc(self, key, base_index)
422 except:
--> 423 raise e
424 return loc, index, index_was_expanded
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/base/tsa_model.py in _get_index_label_loc(self, key, base_index)
411 loc, index, index_was_expanded = (
--> 412 self._get_index_loc(key, base_index))
413 except KeyError as e:
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/base/tsa_model.py in _get_index_loc(self, key, base_index)
363 except (IndexError, ValueError) as e:
--> 364 raise KeyError(str(e))
365 loc = key
KeyError: 'only integers, slices (`:`), ellipsis (`...`), numpy.newaxis (`None`) and integer or boolean arrays are valid indices'
During handling of the above exception, another exception occurred:
KeyError Traceback (most recent call last)
<ipython-input-111-398ffb8cf06c> in <module>
----> 1 pred = model_fit.predict(start=close.index[150], dynamic=False, typ='levels')
~/anaconda3/lib/python3.7/site-packages/statsmodels/base/wrapper.py in wrapper(self, *args, **kwargs)
93 obj = data.wrap_output(func(results, *args, **kwargs), how[0], how[1:])
94 elif how:
---> 95 obj = data.wrap_output(func(results, *args, **kwargs), how)
96 return obj
97
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/arima_model.py in predict(self, start, end, exog, typ, dynamic)
1823 def predict(self, start=None, end=None, exog=None, typ='linear',
1824 dynamic=False):
-> 1825 return self.model.predict(self.params, start, end, exog, typ, dynamic)
1826 predict.__doc__ = _arima_results_predict
1827
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/arima_model.py in predict(self, params, start, end, exog, typ, dynamic)
1196 if not dynamic:
1197 predict = super(ARIMA, self).predict(params, start, end, exog,
-> 1198 dynamic)
1199
1200 start, end, out_of_sample, _ = (
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/arima_model.py in predict(self, params, start, end, exog, dynamic)
717 # will return an index of a date
718 start, end, out_of_sample, _ = (
--> 719 self._get_prediction_index(start, end, dynamic))
720
721 if out_of_sample and (exog is None and self.k_exog > 0):
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/arima_model.py in _get_prediction_index(self, start, end, dynamic, index)
1059
1060 start, end, out_of_sample, prediction_index = (
-> 1061 super(ARIMA, self)._get_prediction_index(start, end, index))
1062
1063 # From _get_predict_end
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/arima_model.py in _get_prediction_index(self, start, end, dynamic, index)
649
650 start, end, out_of_sample, prediction_index = (
--> 651 super(ARMA, self)._get_prediction_index(start, end, index))
652
653 # This replaces the _validate() call
~/anaconda3/lib/python3.7/site-packages/statsmodels/tsa/base/tsa_model.py in _get_prediction_index(self, start, end, index, silent)
477 start, start_index, start_oos = self._get_index_label_loc(start)
478 except KeyError:
--> 479 raise KeyError('The `start` argument could not be matched to a'
480 ' location related to the index of the data.')
481 if end is None:
KeyError: 'The `start` argument could not be matched to a location related to the index of the data.'