example_ode_model()
.$diagnose()
method for the OdeModel
class.diagnose
and stansummary
in the
OdeModelMCMC
class instances.$reliability()
method for the OdeModelMCMC
class. Adds a
recompute_loglik
argument for it.stan_transform
s
that both depend on the ODE solution AND have effect on model likelihood.
This is because transformed parameters block is created so that
the ODE solution is always second last and log likelihood last.plot_r_eff()
which plotted the wrong metric.$reliability()
method for the OdeModelMCMC
class so that
it only ever compares OdeModelGQ
objects due to possible
i/o information loss.plot_metric()
and other plotting functions.$sample_manyconf()
method of OdeModel
class.compute_reliability_metrics()
function.$reliability()
method for the
OdeModelMCMC
class.OdeModelFit
s.y0
always in the same format,
no matter if it is a parameter, data, or transformation.lynxhare
data.$simulate()
method to $gqs()
and implement it also for
OdeModel
.$make_params()
method for OdeModel
.data vector y0
to vector y0
in ODE solver function signatures
so it is possible to have initial state be parameter.name
argument to example_ode_model()
and adds the tmdd
example.stan_param()
and stan_transform()
less vulnerable
to errors.ode_model()
, it is now possible to create also a model where
odefun_vars
and loglik_vars
are empty lists.ode_model()
for creating models.stan_var()
, stan_param()
etc. for model specification.rk45()
, bdf()
, midpoint()
, rk4()
etc. for solver
specification.OdeModel
and OdeModelMCMC
and OdeModelGQ
R6 classes.$sample()
method of the
OdeModel
class.$simulate()
method of OdeModelMCMC
class)OdeModelMCMC
and OdeModelGQ
classes are for example $extract_unflattened()
, $extract_odesol()
,
$plot_odesol()
, $dim()
and $dim_odesol()
.OdeModelMCMC
and
OdeModelGQ
objects, such as log_ratios()
, psis()
,
max_abs_odesol_diff()
and max_abs_loglik_diff()
.Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.