Dear ComRades,
Package CatDyn 1.0-3 is now available on CRAN.
The package allows the estimation of the absolute abundance of a wild
population during a capture season using a nonlinear and recursive
discrete-time structural model.
The response variable is the catch, assumed a random variable produced by
either a multiplicative stochastic process (lognormal distribution) or an
additive stochastic process (normal distribution).
The predictor variables are capture effort, assumed observed exactly, and stock
abundance, a latent predictor.
The data are high frequency (e.g. daily) observations of catch and capture
effort during the season.
The structural model comes in five varieties. The simplest model assumes that
all individuals that can be captured during the season were available at the
start of the season, thus during the season the abundance always decreases (a
pure depletion model), due to natural mortality and capture mortality. The
other versions include one, two, three or four positive perturbations, such as
immigration of recruits, that occur at certain time steps during the season.
The models also includes nonlinear (power) relations between response and
predictors. The models can be fit with natural mortality fixed at a given value
or letting this parameter be estimated as well. Taking into account structural
and stochastic options, CatDyn can fit up to 20 model versions to the same data.
CatDyn uses the recently released optimx package for maximum flexibility in
selecting and using numerical methods implemented in R. CatDyn also has
analytical gradients but in the current version these gradients are not yet
passed to the optimizer; instead they are computed after convergence using
numerical gradients, in order to compare analytical versus numerical gradients
at the maximum likelihood estimates using various methods.
CatDyn estimates all parameters in the log scale and uses function
deltamethod() from package msm to return back-transformed estimates vectors and
covariance matrices.
CatDyn has been used already to estimate abundance of several fish and
invertebrate stocks harvested by industrial and artisanal fishing fleets over
the world.
A manuscript is under preparation with the first application (a squid stock
from the Falkland Islands) and will be submitted to a marine science journal.
Dr. Rubén H. Roa-Ureta,
AZTI Tecnalia, Txatxarramendi Ugartea z/g,
Sukarrieta, Bizkaia, SPAIN
[[alternative HTML version deleted]]
_______________________________________________
R-packages mailing list
r-packa...@r-project.org
https://stat.ethz.ch/mailman/listinfo/r-packages
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
