You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
thanks for creating the package. I also really like the example notebooks. I looked into the notebook: Interface to statsmodels: state space time series models and really like the idea of calculating the difference between the solution of a process-based model (like the logistic growth model) and the data and calcuate the likelihood with a local level state space model (with process and observarional error parameters). I think it is in between a full state space approach with one step ahead prediction of the process-based model and directly updating the states and more simpler aprroaches like the AR1 likelihood that you also described.
My question is: Do you know any scientific articles that use this approach of calculting the likelihood with local level state space model from the difference of the solution of a process-based model and data? Maybe, some references could be linked in the notebook? This would be highly useful for me, because I could cite these articles for justifying the approach in my paper.
Best,
Felix
The text was updated successfully, but these errors were encountered:
Hello,
thanks for creating the package. I also really like the example notebooks. I looked into the notebook: Interface to statsmodels: state space time series models and really like the idea of calculating the difference between the solution of a process-based model (like the logistic growth model) and the data and calcuate the likelihood with a local level state space model (with process and observarional error parameters). I think it is in between a full state space approach with one step ahead prediction of the process-based model and directly updating the states and more simpler aprroaches like the AR1 likelihood that you also described.
My question is: Do you know any scientific articles that use this approach of calculting the likelihood with local level state space model from the difference of the solution of a process-based model and data? Maybe, some references could be linked in the notebook? This would be highly useful for me, because I could cite these articles for justifying the approach in my paper.
Best,
Felix
The text was updated successfully, but these errors were encountered: