ritvikmath / time-series-analysis Goto Github PK

Hi @ritvikmath,
First, your work provided for the community is far more than excellent & super valuable.

In SARIMA Model.ipynb the workaround to improve on the residuals might be problematic:
the original task is a 6-item suffix forecast but the relaxation of that to the average of 1-step forecasts is not really focusing on the same task IMHO.

Hi!

Can you please explain me, why the p values are 0 beyond the third lag for the grangercausalitytest? I mean, they should not be significant, right?

Best regards

Hi,

I am trying your code but I have encountered an error without any modification.

Thanks.

I am quite new in Python ...

When I run your ARMA-Code in Jupyter Notebook, I get an error after block [9] (i.e. acf_vals = acf(first_diff) .....):

NameError Traceback (most recent call last)
in
1 acf_vals = acf(first_diff)
----> 2 plt.bar(range(num_lags), acf_vals[:num_lags])

NameError: name 'num_lags' is not defined

May be you could check it.

Thank you,
Christian