convergence_info
, opt_obj
, and opt_par
provide basic access to the underlying optimization objects, which should be enough for people to construct their own convergence diagnosticsbbmle
settings can get stuck on parameters that cause flat dynamics for simple models -- not sure how important this is for more complex models)bbmle::confint.mle2
, TMB::sdreport
, rstan::summary
SEIR * vax_status
)n
days, and so the lagged differencing needs to account for variable n
Naming conventions for value vector? troubleshoot test_calibrate more docs for new params_timevar structure issue/document testLevel stuff in README
update method for calibrations
names? Symbol.Date values -> time_params
we need to get importance weighting going etc. (or MCMC, but imp weighting should work if anything will)
instabilities etc.
simulate from simple model
simulate from spline model
test spline variations: number of spline parameters? effects of var choice on knot placement? effects of (strong) penalization on spline coeffs?
document !!!!
pop_pred_samp
Makiness:
add testing intensity?
get_r
work for r<1 ...calibrate
calibrate
/mle_fun
: better error handling if dates/vars don't matchcalibrate
substitute mle2 for optim?forecast_ensemble
?process error version
transition TODO list to GH issues??? (items marked Z have been posted)
beta*(S/N)^het_alpha*I
do_hazard
option and switch to this as defaultupdate_foi
functionwrite_params
, read_params
; keep labels, unevaluated values, etc. as attributes, get start date as attribute. (Store as separate column, or commented row, or ... ???) [store params as JSON???]testing_flow.md
)Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.