bsm: Fit a Bayesian Selectivity Model
In tmelliott/bsm: Bayesian Selectivity Modelling of Fisheries Trawl Data

Description Usage Arguments Details Value Methods (by generic) Author(s)

Creates and fits a selection curve to the supplied bsmdata object. The correct data list is created, the JAGS model constructed, and the model is run with default settings. It is also possible to suppress the running of the JAGS and simply output the model and data.

bsm(x, family = "binomial", curve = "logistic", check.od = TRUE,
  od = FALSE, combine = FALSE, random = NULL, L50 = ~1, SR = ~1,
  phi = if (attr(x, "paired")) ~1 else NULL, delta = if (curve ==
  "richards") ~1 else NULL, priors = NULL, inits = NULL,
  length.dist = "iid", parameters = par$summary.parameters,
  extra.pars = NULL, file = NULL, fit = TRUE, include.lambda = FALSE,
  n.samples = 1000, n.thin = 1, n.burn = n.samples * n.thin,
  n.chains = 3, max.attempts = 10, quiet = FALSE, progress.bar = "text",
  show.info = TRUE, ...)

## S3 method for class 'bsmfit'
print(x, model = FALSE, coda = FALSE, ...)

## S3 method for class 'bsmfit'
summary(object, p.values = FALSE, formula.values = FALSE,
  predict.values = NULL, ...)

`x`	an object of class `bsmdata`, from the `bsmData` function
`family`	`"logistic"` (default) or `"poisson"`, the family for the likelihood
`curve`	`"logistic"` or `"richards"`, the type of selection curve to be fitted (will likely allow other options, for example "Bspline", future)
`check.od`	logical, if `TRUE`, then overdispersion estimates will be produced
`od`	logical, if `TRUE`, the model will be fit allowing for overdispersion
`combine`	logical, if `TRUE`, then the 'combined hauls' approach is used, otherwise a hierarchical approach is used.
`random`	vector of parameters to have hierarchical or random effects
`L50`	the formula for L50, default is L50 = ~1
`SR`	the formula for SR, default is SR = ~1
`phi`	the formula for phi, default is phi = ~1
`delta`	the formula for delta, default is delta = ~1
`priors`	the prior distributions for specified parameters. See details.
`inits`	initial values for parameters
`length.dist`	'iid' or "multinomial", the type of length distribution for the lambda parameters. See details for more information.
`parameters`	the parameters to save in the JAGS output. NOTE: only use this if you really really only want these parameters - may cause errors in summary output and plots depending on parameters selected.
`extra.pars`	additional parameters to be tracked as well as `parameters`. Lets you keep the default parameters.
`file`	the file name to save the JAGS model in. If `NULL`, then a temporary file is created
`fit`	logical, if `TRUE` then the JAGS model is fit, otherwise only the model and data are returned.
`include.lambda`	logical, if `TRUE`, the lambda parameter will be saved as well
`n.samples`	the total number of samples to obtain (excludes burnin and thin)
`n.thin`	the thinning interval
`n.burn`	the burn in period
`n.chains`	the number of chains
`max.attempts`	the maximum number of times bsm will attempt to fit the JAGS model (often due to bad initial values, the model cannot be initiated and fails)
`quiet`	suppress all output messages from JAGS?
`progress.bar`	one of "none", "text" or "gui" for the JAGS progress bar
`show.info`	logical, if `FALSE`, the package information will not be printed
`...`	additional arguments
`model`	logical, if `TRUE`, then the JAGS model is printed, otherwise summary information is provided
`coda`	logical, if `TRUE`, then the summary from the `coda` package is used
`object`	a bsmfit object
`p.values`	logical, include significance tests for parameters != 0?
`formula.values`	logical, display values in formulae?
`predict.values`	values to be included in the prediction