The 365 Data Science team is proud to invite you to our own community forum. A very well built system to support your queries, questions and give the chance to show your knowledge and help others in their path of becoming Data Science specialists.
Ask
Anybody can ask a question
Answer
Anybody can answer
Vote
The best answers are voted up and moderated by our team

ARMA model

ARMA model

0
Votes
1
Answer

Hi!
I had an error in the 3 lecture of ARMA in time-series. When I fit the model it says:’The computed initial AR coefficients are not stationary You should induce stationarity, choose a different model order, or you can pass your own start_params’.
I guess it’s not something about the code or the data
model_returns_arma_33=ARMA(train.returns[1:],order=(3,3))
results_returns_arma_33=model_returns_arma_33.fit()

I also did adfuller test, which has shown very low p-value.

sts.adfuller(train.returns[1:])

(-17.03445719098116, 8.28053702031718e-30, 17, 5002, {‘1%’: -3.431658008603046, ‘5%’: -2.862117998412982, ‘10%’: -2.567077669247375}, 16035.926219345134)

1 Answer

365 Team
0
Votes

Hi Max, 
 
Thanks for reaching out!
 
When this occurs, you need to specify the number of starting AR parameters you’re giving the model before fitting. Usually, 1 more than the number of lags you’re using is sufficient, but you might need to add more depending on the data. 
 
Hence, just add start_ar_lags = 4 to the fit function, like so: 

results_returns_arma_33=model_returns_arma_33.fit(start_ar_lags = 4)

 
Hope this helps!
Best,
365 Vik