While backtesting a kalman filter model, I hit a performance bottleneck that raised the following question:

Given that all the input data is available for a time range, is it possible to evaluate the end result of the model without calculating each interim step?

While my hunch is that it isn't possible, I wasn't able to find a definitive answer online.


put on hold as unclear what you're asking by Taylor, Michael Chernick, mkt, kjetil b halvorsen, jpmuc 2 days ago

Please clarify your specific problem or add additional details to highlight exactly what you need. As it's currently written, it’s hard to tell exactly what you're asking. See the How to Ask page for help clarifying this question. If this question can be reworded to fit the rules in the help center, please edit the question.

  • $\begingroup$ Doesn't this just amount to finding a more efficient model, or of calculating a model? If you know the end result, why would you do additional calculations? $\endgroup$ – mkt Feb 12 at 12:15
  • $\begingroup$ Yes, it amounts to finding a more efficient model. Generally the kalman filter is used as an online model so the recursive calculations are necessary. However, I'm wondering if it's possible to obtain the end result without doing the interim calculations - given that all the data is available $\endgroup$ – bphi Feb 12 at 13:58