secrgam-package: SECR with Density Surface GAMs

In david-borchers/secrgam: Spatially Explicit Capture-Recapture with Generalised Additive Density Models

Description

Implements Genaralized Additive Models for modelling density surfaces for Spatially Explicit Capture-Recapture, by means of regression splines.

Details

Package:	secrgam
Type:	Package
Version:	1.0
Date:	2014-01-12
License:	GNU General Public License Version 2 or later

Package secrgam uses regression splines to implement flexible density surface estimation for Spatially Explicit Capture-Recapture methods, using maximum likelihood inference. At its core is package secr, which does all the actual estimation; package secrgam is just a set of wrapper functions and some plotting functions that set up the regression spline bases and then call secr.

The package currently allows spline models to be specified using syntax from package mgcv using the s and te functions. The gam function is used with fit=FALSE to obtain regression spline design matrices.

Author(s)

David Borchers [e-mail]

Darren Kidney [e-mail]

Maintainer: David Borchers [e-mail]

See Also

secr

Examples

library(secrgam)

data(Boland.leopards1)
data(Boland.alt.image)

# plot survey region and traps
Boland.cameras = traps(Boland.CH1)
image(Boland.alt.image, col = terrain.colors(60),main="Altitude map")
contour(Boland.alt.image, add = TRUE)
plot(Boland.cameras, add = TRUE, detpar = list(pch = "+", cex = 1.2))

# summarise and plot capture histories
summary(Boland.CH1)
plot(Boland.CH1, border = 0, rad = 0.01, tracks = TRUE, icolour = colors()[seq(2, 202, 10)])

# make a model with dependence on altitude via smooth with 4 degrees of freedom
model = list(D ~ s(alt, k = 4), g0 ~ 1, sigma ~ 1)

# fit model
## Not run: 
fit <- secrgam.fit(capthist = Boland.CH1, model = model, mask = Boland.mask1, trace = FALSE)
fit # look at fit results

# plot fitted surface:
plot(fit,asp=1)
plot(traps(Boland.CH1), add = TRUE)

# plot smooths:
plot(fit, type = "smooth")

## End(Not run)

david-borchers/secrgam documentation built on May 14, 2019, 9:30 p.m.