makePostSurvivalObjs: Make list of survival objects corresponding to posterior of Bayes fit

Description

Model takes a data-frame and fits a Bayesian model using the chosen explanatory variable, and the package MCMCglmm, fitting an error variance of 1; then create the resulting list of survival objects necessary to get Bayes posterior of IPM outputs

Usage

makePostSurvivalObjs(dataf, explanatoryVariables="size+size2",
meanB = rep(0, 3), varB = rep(1e+10),burnin=3000, nitt = 50000)

Arguments

dataf

a dataframe with columns `size',`sizeNext', and `surv' (`size' is size at t, `sizeNext' is size at t+1, and `surv' is 0s and 1s, indicating if the individual survived or not); facultatively, dataf may include `covariate' (discrete or continuous) or ot

explanatoryVariables

a formula describing the desired explanatory variables (interactions, etc) in classical R style, i.e. separated by `+', `*', `:'. Possible covariates include `size', 'size2' (size^2), `size3' (size^3),`logsize' (log(size)), `logsize2' (log(size)^2), and `

meanB

vector defining the means of the priors of the coefficients for survival (should be the same length as desired coeff). CURRENTLY NOT IMPLEMENTED.

varB

vector defining the variance of the prior of the coefficients for survival,as for the mean. CURRENTLY NOT IMPLEMENTED.

burnin

numeric defining burn in for MCMCglmm

nitt

numeric defining the number of iterations

Value

a list of survObjOverDisp survival objects, of length nitt; appropriate survival methods are defined that use Diggle et al. (2004)'s correction for estimating the expectation of survival following non-linear transform of a distribution on the logit scale. } references{ Diggle, P., P. Heagerty, K. Liang, and S. Zeger. 2004. Analysis of Longitudinal Data. 2nd edition. Oxford University Press.%% ~put references to the literature/web site here ~} author{ C. Jessica E. Metcalf, Sean M. McMahon, Roberto Salguero-Gomez & Eelke Jongejans }
%% ~Make other sections like Warning with \section{Warning }{....} ~ seealso{ code{makePostGrowthObjs}, code{makeListPmatrix} %% ~~objects to See Also as \code{\link{help}}, ~~~} examples{ dff <- generateData() svlist <- makePostSurvivalObjs(dataf=dff, explanatoryVariables="size+size2+covariate", meanB = rep(0, 3), varB = rep(1e+10), burnin = 100, nitt = 500)
plot(1:20,surv(1:20,data.frame(covariate=as.factor(1)),svlist[[1]]), type="n", xlab="size", ylab="survival") for (k in 1:length(svlist)) points(1:20,surv(1:20,data.frame(covariate=as.factor(1)),svlist[[k]]), type="l",col="grey")
}

Details

Note that the use of priors is not yet implemented, and the function does not return output of Bayes analysis which might be useful for assessment; in development. Error variance is set to one. See Course Notes from MCMCpack for reasons that this results in slightly different parameters that would be obtained using a glm with family=binomial.