initiate.startValues: The function that initiates starting values for a single chain.

Description

The function initiates starting values for a single chain. Users are allowed to set some non-null values to starting values for a set of parameters. The function will automatically generate starting values for any parameters whose values are not specified.

Usage

initiate.startValues(Y, lin.pred, data, model, cluster = NULL, 
                   beta1 = NULL, beta2 = NULL, beta3 = NULL, beta = NULL,
                   gamma.ji = NULL, theta = NULL,
                   V.j1 = NULL, V.j2 = NULL, V.j3 = NULL, V.j = NULL,
                   WB.alpha = NULL, WB.kappa = NULL, 
                   MVN.SigmaV = NULL, Normal.zeta = NULL, 
                   DPM.class = NULL, DPM.tau = NULL)

Arguments

For BayesID, it is a data.frame containing semi-competing risks outcomes from n subjects. For BayesSurv, it is a data.frame containing univariate time-to-event outcomes from n subjects.

lin.pred

For BayesID, it is a list containing three formula objects that correspond to $h_g()$, $g$=1,2,3. For BayesSurv, it is a formula object that corresponds to $h()$.

data

a data.frame in which to interpret the variables named in the formula(s) in lin.pred.

model

a character vector that specifies the type of components in a model. Check BayesID and BayesSurv.

cluster

a vector of cluster information for n subjects. The cluster membership must be set to consecutive positive integers, $1:J$.

beta1

starting values of $\beta_1$ for BayesID.

beta2

starting values of $\beta_2$ for BayesID.

beta3

starting values of $\beta_3$ for BayesID.

beta

starting values of $\beta$ for BayesSurv.

gamma.ji

starting values of $\gamma$ for BayesID.

theta

starting values of $\theta$ for BayesID.

V.j1

starting values of $V_{j1}$ for BayesID.

V.j2

starting values of $V_{j2}$ for BayesID.

V.j3

starting values of $V_{j3}$ for BayesID.

V.j

starting values of $V_{j}$ for BayesSurv.

WB.alpha

starting values of the Weibull parameters, $\alpha_g$ for BayesID. starting values of the Weibull parameter, $\alpha$ for BayesSurv.

WB.kappa

starting values of the Weibull parameters, $\kappa_g$ for BayesID. starting values of the Weibull parameter, $\kappa$ for BayesSurv.

MVN.SigmaV

starting values of $\Sigma_V$ in DPM models for BayesID.

Normal.zeta

starting values of $\zeta$ in DPM models for BayesSurv.

DPM.class

starting values of the class membership in DPM models for BayesID and BayesSurv.

DPM.tau

starting values of $\tau$ in DPM models for BayesID and BayesSurv.

Value

initiate.startValues returns a list containing starting values for a sigle chain that can be used for BayesID and BayesSurv.

References

Lee, K. H., Haneuse, S., Schrag, D., and Dominici, F. (2015), Bayesian semiparametric analysis of semicompeting risks data: investigating hospital readmission after a pancreatic cancer diagnosis, Journal of the Royal Statistical Society: Series C, 64, 2, 253-273. Lee, K. H., Dominici, F., Schrag, D., and Haneuse, S., Hierarchical models for cluster-correlated semicompeting risks data with application to quality of end-of-life care for pancreatic cancer, arXiv:1502.00526; submitted.

Examples

Run this code

## See Examples in \code{\link{BayesID}} and \code{\link{BayesSurv}}.