I am using ARIMA to forecast the time series of some medical data. I was wondering if I can take the ARIMA model I fit to my data and get some numbers that describe just the trend and seasonality separately. This would be useful for me because it would allow me to see what my model's trend rate is without seasonality affecting the results. Please let me know if you have any questions. Thanks.
I was unable to find anything from a google search and have idea where to start. I looked into seasonal decompose but that seems to get trend and seasonality of my actual data, not the model fit to the data.
I think that if you want to know the trend and the seasonality of your model you should first make prediction on a large range of date using .forecast(bignumber). Then on this prediction you could do decomposition using statsmodels.tsa.seasonal.seasonal_decompose. Like that you will have a clear idea of the trend and the seasonality learned by your ARIMA model. After, if you want to estimate the expression of your trend you can train a linear or polynomial model on the trend decomposed.