fit_vb | R Documentation |
For use with data for a single species.
fit_vb(
dat,
sex = c("female", "male", "all"),
method = c("tmb", "mpd", "mcmc"),
downsample = Inf,
chains = 4L,
iter = 1000L,
cores = parallel::detectCores(),
allow_slow_mcmc = FALSE,
est_method = median,
min_samples = 50L,
too_high_quantile = 1,
uniform_priors = FALSE,
ageing_method_codes = NULL,
usability_codes = c(0, 1, 2, 6),
check_convergence_tmb = TRUE,
tmb_inits = list(k = 0.5, linf = 40, log_sigma = log(0.1), t0 = -1),
...
)
dat |
Input data frame. Should be from |
sex |
Either "male" or "female". |
method |
|
downsample |
If not |
chains |
Number of Stan chains. |
iter |
Number of Stan sampling iterations. |
cores |
Number of cores for Stan. |
allow_slow_mcmc |
Logical. If |
est_method |
If MCMC this defines how to summarize the posterior. Should
be a function such as |
min_samples |
The minimum number of fish before a model will be fit. |
too_high_quantile |
A quantile above which to discard weights and lengths. Can be useful for outliers. Defaults to including all data. |
uniform_priors |
Logical. If true then uniform priors will be used. |
ageing_method_codes |
A numeric vector of ageing method codes to filter
on. Defaults to |
usability_codes |
An optional vector of usability codes.
All usability codes not in this vector will be omitted.
Set to |
check_convergence_tmb |
Logical. |
tmb_inits |
A named list of initial parameter values for the TMB model. |
... |
Any other arguments to pass on to |
## Not run:
# with `rstan::optimizing()` for the mode of the posterior density:
model_f <- fit_vb(pop_samples, sex = "female")
model_m <- fit_vb(pop_samples, sex = "male")
plot_vb(model_f, model_m)
model_f$model
model_f$predictions
# You can also fit both sexes combined if you want.
# Just note that you need to specify the colours explicitly in the plot.
model_all <- fit_vb(pop_samples, sex = "all")
plot_vb(object_all = model_all, col = c("All" = "black"))
# with MCMC via Stan (slower):
x <- fit_vb(pop_samples, method = "mcmc",
chains = 1, iter = 800, seed = 123) # just for a fast example
x$pars
x$predictions
x$data
x$model
posterior <- rstan::extract(x$model)
hist(posterior$linf)
# If less than `min_samples`, fit_vb() returns an empty object that
# plot_vb() will correctly parse and produce an empty plot:
obj <- fit_vb(pop_samples[1:2,])
plot_vb(obj, obj)
## End(Not run)
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.