Fit catch data to sizeselectivity models
Description
First fits selection curves by loglinear fitting using
glm
and then (if tangle is TRUE) uses the
SELECT method (of Millar 1992) to fit a slightly modified
model with a tangle parameter (see Bromaghin 2005 for the
concept of tangle parameters).
Usage
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24  fit.catch(catch, sel.curve = norm.loc, tangle = TRUE,
perimeter.factor = 1, tol = 1e08, omega0 = 0.1, effort = NULL)
## S3 method for class 'fit.catch'
AIC(object, corrected = TRUE, ..., k = 2)
## S3 method for class 'fit.catch'
coef(object, ...)
## S3 method for class 'fit.catch'
coefficients(object, ...)
## S3 method for class 'fit.catch'
print(x, digits = max(3, getOption("digits")  3), ...)
## S3 method for class 'fit.catch'
plot(x, y, xlab = "Fork length (mm)",
ylab = "Relative selection probability", resolution = 500,
plot.type = c("selcurve", "totalselcurve", "residuals",
"observedvrsexpected"), max.cex = 2, min.cex = 0.5, leg.pos = "topleft",
data.name = NULL, justtangle = TRUE, ...)
## S3 method for class 'fit.catch'
deviance(object, ...)

Arguments
catch 
an object of class 
sel.curve 
one of a number of selection curve function including norm.loc, norm, lognorm, gamm and inv.gau. These functions create a range of functions that are required to make various calculations. In effect, the selection curve chosen defines the model to be fitted. 
tangle 
If TRUE, a tangle parameter is included in the fitted process (Bromaghin (2005). 
perimeter.factor 
Factor by which to multiply the inputted mesh sizes to obtain mesh perimeters, which are required by the analyses. perimeter.factor = 4; if using stretch mesh perimeter.factor = 2, the default). 
x 
an object of class 
object 
an object of class 
y 
not used. only for consistency with S3 plot method. 
xlab 
xaxis label 
ylab 
yaxis label 
resolution 
higher numbers make smoother selection curves 
plot.type 
type of plot to produce, either "selcurve", "totalselcurve", or "residuals" for selection curves, total selection curve, or residual plot; in this last type, the sizes of the circles are proportional to deviance 
max.cex 
maximum size of the deviance circles 
min.cex 
minimum size of the deviance circles 
leg.pos 
position of the legend if necessary, see

data.name 
name of the data set to be plotted 
justtangle 
should only the tangle model be plotted? 
corrected 
if TRUE, the smallsamplesize corrected version of AIC is returned (recommended) 
k 
not used. only for consistency with default S3 method. 
tol 
tolerance used to decide on whether fitted values are numerically zero. used to correct AIC values. this correction depends on sample size, and the standard way to calculate sample size is as the number of mesh sizes times the number of fish length categories and then to subtract from this number the number of zero fitted counts 
omega0 
initial value for the tangle parameter 
digits 
number of digits to round to 
... 
additional arguments to be passed 
effort 
TODO (sorry!) 
Value
An object of class fit.catch
with components:
catch the original catch
object
sel.curve the original selection curve
theta the fitted parameters of the selection curve
mu the fitted values
res the deviance residuals (see Millar and Fryer 1999)
dev model deviance(s)
tangle TRUE if tangle parameter fitted, FALSE otherwise
call the matched call
l the 'length matrix'.
m the 'mesh matrix' (with input mesh sizes multiplied by
perimeter.factor
).
Author(s)
Steve Walker
References
J.F. Bromaghin (2005) A versatile net selectivity model, with application to Pacific salmon and freshwater species of the Yukon River, Alaska. Fisheries Research 74: 157168.
R.B. Millar & R.J. Fryer (1999) Estimating the sizeselection curves of towed gears, traps, nets and hooks. Reviews in Fish Biology and Fisheries 9: 89116.
Examples
1 2 3 4 5  data(north.pike)
np < make.catch(north.pike$MESH, north.pike$FLEN)
plot(np)
(np.fit < fit.catch(np, gamm))
plot(np.fit)
