Search code examples
pythonstatsmodelsarima

statsmodels SARIMAX - How to access parameters by name?

I am fitting an ARMA model with a linear trend to a sequence and I would like to access the parameters after optimization to make some test. It seems like I can access the parameters via the attribute params, but it's just an array with all the parameters in it. Is there a more convenient way to access the parameters (e.g. a dictionary)?

import numpy
from statsmodels.tsa.statespace.sarimax import SARIMAX

N_TRAIN = 100

train_data = numpy.random.randn(N_TRAIN)

model = SARIMAX(
    endog=train_data,
    order=(2, 0, 2),                
    seasonal_order=(0, 0, 0, 0),    
    trend='ct',                     
    enforce_stationarity=False,
    enforce_invertibility=False
)

fitted_model = model.fit(disp=False, maxiter=100, method='powell')

print(fitted_model.params)

Out:
[-0.03781736 -0.00285008  0.88970682 -0.88314944 -1.05743719  0.89770456
  0.84405409]
Solution

  • If the input endog array is a Numpy array, then all results from Statsmodels are also Numpy arrays.

    If you make your data into a Pandas series, then the returned params object will be a Pandas Series including parameter names.

    For example you could do: train_data = pd.Series(numpy.random.randn(N_TRAIN)) (after doing import pandas as pd).