This function can handle nests with unknown nest age, nest-specific covariates(both discrete and continues covariates), and any hazard rate function(or equivelently survival rate function) as long as it is a smooth function.
Usage
nestsr(jj, nx, nn, ul, ur, zl, zr, x, y, a, b, sigma, day, enc, covar, n0, ntotal)
Arguments
jj
The number of time units(days) a nest is required to survive to be considered successful.
nx
The number of covariates.
nn
The number of observed nests (sample size).
ul
The youngest possible age that the nest could have been when first encounterd.
ur
The oldest possible age that the nest could have been when first encounterd.
zl
The smallest possible number of time units from the first encounter date to the outcome date.
zr
The largest possible number of time units from the first encounter date to the outcome date.
x
The nest-specific covariate-matrix.
y
The nest fate.
a
The specified value for hyperparameter a. The prior gamma(a,b) is for age effect variances.
b
The specified value for hyperparameter b. The prior gamma(a,b) is for age effect variances.
sigma
The specified values for hyperparameters age-effect variances.
day
The iniital values for age effect of outcome rates.
enc
The initial values for age effect of encounter rates.
covar
The initial values for the coefficients for covariates.
n0
The number of burn-in cycles.
ntotal
The number of total Gibbs cycles.
Details
The Bayesian estimate of parameter is computed from its posterior distribution which is simulated by Gibbs sampler. Users need to specify a set of initial values ,the number of burn-in cycles and the total number of Gibbs sampling cycles.
Value
The BEANSP returns the esitmate and corresponding standard deviation for all key parameters, the average age-specific survival rate and average cumulative survival rate (average over all nests), and selected age-specific survival rate and cumulative survival rate for individual nest. It also outpus a model selection criterion DIC (Spiegelhalter et al. 2002).
jj
nest period time.
enc
numerical values of estimate of encounter age effect for all age.
day
numerical values of estimate of outcome age effect for all age .
sigma
numerical values of estimate of age effect variances.
covar
numerical values of estimate of regression coefficients.
q
numerical values of estimate of age-specific outcome rates.
del
numerical values of estimate of age-specific encounter rates.
sr
numerical values of estimate of individual age-specific survival rates.
asr
numerical values of estimate of average age-specific survival rates.
casr
numerical values of estimate of average cumulative age-specific survival rates.
DIC
model selection criterion DIC value
Dbar
a expectation measure of how well the model fits the data.
pd
a measure for the effective number of parameters of the model.
trace1
numerical trace values for encounter age effect for all age.
trace2
numerical trace values for outcome age effect for all age.
trace3
numerical trace values for regression coefficients.
trace4
numerical trace values for age effect variances.
venc
standard deviation of encounter age effect for all age.
vday
standard deviation of outcome age effect for all age.
vsigma
standard deviation of age effect variances.
vcovar
standard deviation of regression coefficients.
vq
standard deviation of age-specific outcome rates.
vdel
standard deviation of age-specific outcome rates.
vsr
standard deviation of individual age-specific survival rates.
vasr
standard deviation of average age-specific survival rates.
vcasr
standard deviation of average cumulative age-specific survival rates.
Author(s)
Chong He, Yiqun Yang, Jing Cao
References
Cao, J., He, C., Suedkamp Wells, K.M., Millspaugh, J.J., and Ryan, M.R. (2009). Modeling age and nest-specific survival using a hierarchical Bayesian approach. Biometrics, 65, 1052-1062.