Dates in timeseries models¶

In [ ]:

from __future__ import print_function
import statsmodels.api as sm
import numpy as np
import pandas as pd

Getting started¶

In [ ]:

data = sm.datasets.sunspots.load()

Right now an annual date series must be datetimes at the end of the year.

In [ ]:

from datetime import datetime
dates = sm.tsa.datetools.dates_from_range('1700', length=len(data.endog))

Using Pandas¶

Make a pandas TimeSeries or DataFrame

In [ ]:

endog = pd.TimeSeries(data.endog, index=dates)

Instantiate the model

In [ ]:

ar_model = sm.tsa.AR(endog, freq='A')
pandas_ar_res = ar_model.fit(maxlag=9, method='mle', disp=-1)

Out-of-sample prediction

In [ ]:

pred = pandas_ar_res.predict(start='2005', end='2015')
print(pred)

Using explicit dates¶

In [ ]:

ar_model = sm.tsa.AR(data.endog, dates=dates, freq='A')
ar_res = ar_model.fit(maxlag=9, method='mle', disp=-1)
pred = ar_res.predict(start='2005', end='2015')
print(pred)

2005-12-31    20.003285
2006-12-31    24.703979
2007-12-31    20.026123
2008-12-31    23.473638
2009-12-31    30.858572
2010-12-31    61.335449
2011-12-31    87.024691
2012-12-31    91.321256
2013-12-31    79.921629
2014-12-31    60.799526
2015-12-31    40.374879
Freq: A-DEC, dtype: float64

This just returns a regular array, but since the model has date information attached, you can get the prediction dates in a roundabout way.

In [ ]:

print(ar_res.data.predict_dates)

[ 20.00328112  24.70398853  20.02613031  23.47364995  30.85857026
  61.33544403  87.02467843  91.32123263  79.92159878  60.79948588
  40.37483539]

Note: This attribute only exists if predict has been called. It holds the dates associated with the last call to predict.