statsmodels.tsa.arima_process.arma2ar

statsmodels.tsa.arima_process.arma2ar(ar, ma, nobs=100)[source]

get the AR representation of an ARMA process

Parameters:

ar : array_like, 1d

auto regressive lag polynomial

ma : array_like, 1d

moving average lag polynomial

nobs : int

number of observations to calculate

Returns:

ar : array, 1d

coefficients of AR lag polynomial with nobs elements

Notes

This is just an alias for ar_representation = arma_impulse_response(ma, ar, nobs=100) which has been fully tested against MATLAB.