pints-team
References in state space notebook
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
@FelixNoessler Good question. I don't know actually of any papers where the model is anything but a simple, non-mechanistic process model. (Could be fertile ground for a research article :) .)
