epiverse-trace / quickfit

[SUSPENDED] Toolbox of model fitting helper functions
https://epiverse-trace.github.io/quickfit/
Other
2 stars 1 forks source link

Possible MCMC parameter estimation #1

Open joshwlambert opened 1 year ago

joshwlambert commented 1 year ago

The package currently contains functionality to estimate the maximum likelihood parameters given a likelihood function. This likelihood function could also be used in a Bayesian estimation procedure using MCMC. There is no plan to develop this functionality unless people request it. If you think MCMC estimation would be useful in {quickfit} please leave a comment in this issue.

Bisaloo commented 1 year ago

Is this covered by https://github.com/Bisaloo/simpleMH? As mentioned in the DESCRIPTION, this is just a proof of concept / for teaching purposes but it could be improved & extended.

joshwlambert commented 1 year ago

Thanks for sharing the link. As you know there are many Bayesian/MCMC fitting packages in R and we don't want to reinvent the wheel.

I think we need a concrete use case before adding anything to {quickfit}. Even if that is a very simple comparison of MLE + CI versus posterior distribution of parameter estimates. We should also recognise that any MCMC implementation will likely add dependencies.