R/sep.2z.R

In zoib: Bayesian Inference for Beta Regression and Zero-or-One Inflated Beta Regression

sep.2z <-
function(y, n, xmu.1, p.xmu, xsum.1, p.xsum, 
                   zdummy, qz,nz0, m, rid, EUID, nEU,
                   prior1, prior2, prior.beta, prior.Sigma, 
                   prec.int, prec.DN, lambda.L1, lambda.L2,lambda.ARD,
                   scale.unif, scale.halft, link, n.chain, inits, seed) 
{
  dataIn <- vector("list",20)
  names(dataIn) <- c("n","y","xmu.1","p.xmu","xsum.1","p.xsum","z",
                     "nz0","qz","m","cumm", "zero","link","prior1",
                     "prior2","hyper","rid","EUID","nEU","hyper2")
  dataIn[[1]] <- n      
  dataIn[[2]] <- y
  dataIn[[3]] <- as.matrix(xmu.1)
  dataIn[[4]] <- p.xmu
  dataIn[[5]] <- as.matrix(xsum.1)
  dataIn[[6]] <- p.xsum      
  dataIn[[7]] <- zdummy
  dataIn[[8]] <- nz0
  dataIn[[9]] <- qz
  dataIn[[10]]<- m
  dataIn[[11]]<- c(0,cumsum(m[-nz0]))   
  dataIn[[12]]<- rep(0,n) 
  dataIn[[13]]<- link
  dataIn[[14]] <- prior1
  dataIn[[15]] <- prior2 
  dataIn[[16]] <- as.matrix(cbind(prec.int,prec.DN,lambda.L1,lambda.L2,lambda.ARD))          
  dataIn[[17]] <- rid
  dataIn[[18]] <- EUID 
  dataIn[[19]] <- nEU
  if(grepl("unif",prior.Sigma))  dataIn[[20]] <- scale.unif
  if(grepl("halfcauchy",prior.Sigma)) dataIn[[20]] <- scale.halft   
  
  if(is.null(seed)){
    init <- function(rngname, rngseed ){
      rho1 <- runif(1,-0.5,0.5)
      rho2 <- runif(1,-0.5,0.5) 
      rho3 <- runif(1, rho1*rho2 - sqrt((1-rho1^2)*(1-rho2^2)), 
                    rho1*rho2 + sqrt((1-rho1^2)*(1-rho2^2)))
      return(
        list("tmp1" = rnorm(1,0,0.1),
             "tmp2" = rnorm(1,0,0.1),
             
             "b.tmp" = matrix(rnorm((p.xmu-1)*4,0,0.1),ncol=4),
             "d.tmp" = matrix(rnorm((p.xsum-1)*4,0,0.1),ncol=4),
             
             "sigmab.L1" = runif((p.xmu-1),0,2), 
             "sigmad.L1" = runif((p.xsum-1),0,2), 
             
             "taub.ARD" = runif((p.xmu-1),0,2), 
             "taud.ARD" = runif((p.xsum-1),0,2), 
             
             "taub.L2" = runif(1,0,2), 
             "taud.L2" = runif(1,0,2),
             
             "sigma.VC1" = runif(nz0,0.25,2),
             "t" = runif(nz0,0.25,1),       
             "scale1" = runif(qz,0.25,2),
             "scale2" = runif(qz,0.25,2),
             
             "rho1" = rho1,
             "rho2" = rho2,
             "rho3" = rho3))}
      inits.internal <- list(init( ));
      if(n.chain >= 2) {
        for(j in 2:n.chain) inits.internal <- c(inits.internal,list(init()))} 
    } else{
      init <- function(rngname, rngseed ){
      rho1 <- runif(1,-0.5,0.5)
      rho2 <- runif(1,-0.5,0.5) 
      rho3 <- runif(1, rho1*rho2 - sqrt((1-rho1^2)*(1-rho2^2)), 
                  rho1*rho2 + sqrt((1-rho1^2)*(1-rho2^2)))
      return(
      list("tmp1" = rnorm(1,0,0.1),
           "tmp2" = rnorm(1,0,0.1),
           
           "b.tmp" = matrix(rnorm((p.xmu-1)*4,0,0.1),ncol=4),
           "d.tmp" = matrix(rnorm((p.xsum-1)*4,0,0.1),ncol=4),
           
           "sigmab.L1" = runif((p.xmu-1),0,2), 
           "sigmad.L1" = runif((p.xsum-1),0,2), 
           
           "taub.ARD" = runif((p.xmu-1),0,2), 
           "taud.ARD" = runif((p.xsum-1),0,2), 
           
           "taub.L2" = runif(1,0,2), 
           "taud.L2" = runif(1,0,2),
           
           "sigma.VC1" = runif(nz0,0.25,2),
           "t" = runif(nz0,0.25,1),       
           "scale1" = runif(qz,0.25,2),
           "scale2" = runif(qz,0.25,2),
           
           "rho1" = rho1,
           "rho2" = rho2,
           "rho3" = rho3,
           
           .RNG.name = rngname, 
           .RNG.seed = rngseed))}    
  
  # 1b, 2d, 
  # 3 SigmaVC (sigma.VC1 or t),SigmaUN (scale1 or scale2),
  # 4 rho1,2,3
  set.seed(seed[1]); inits.internal <- list(init("base::Super-Duper", seed[1]));
  if(n.chain >= 2) {
    for(j in 2:n.chain){ 
      set.seed(seed[j]); 
      inits.internal <- c(inits.internal,list(init("base::Wichmann-Hill",seed[j])))}}  
    }
  
  if(!is.null(inits)){
    
  for(i in 1:n.chain){
    
    if(!is.null(inits[[i]]$b)) {
      inits.internal[[i]][[1]] <- inits[[i]]$b[1]
      if(p.xmu>=2) inits.internal[[i]][[3]] <- matrix(rep(inits[[i]]$b[2:p.xmu,1],4), 
                                         ncol=4, byrow=FALSE)}
    if(!is.null(inits[[i]]$d)) {
      inits.internal[[i]][[2]] <- inits[[i]]$d[1]
      if(p.xsum>=2) inits.internal[[i]][[4]] <- matrix(rep(inits[[i]]$d[2:p.xsum,1],4), 
                                         ncol=4, byrow=FALSE)}
    
    if(!is.null(inits[[i]]$sigma)) {
      inits.internal[[i]][[11]]<- inits[[i]]$sigma
      inits.internal[[i]][[12]]<- inits[[i]]$sigma
      inits.internal[[i]][[13]]<- runif(qz,0.25,2)
      inits.internal[[i]][[14]]<- runif(qz,0.25,2)
    }
    
    # check PD of the initial R matrix
    if(!is.null(inits[[i]]$R)) {
      notuse <-FALSE
      Rele <- inits[[i]]$R
      size <- (sqrt(1+8*length(Rele))-1)/2 # (# of random effects)
      R <- diag(size)
      R[upper.tri(R, diag=TRUE)] <- Rele 
      R <- R + t(R) - diag(diag(R))
      pd <- all(eigen(R)$values>0)
      if(!pd) {
        notuse <- TRUE
        warning('the specified initial correlation matrix is not positive definite')
        warning('Internal initial value are used')
        break}
      else{
        if(size==2) inits.internal[[i]][[15]] <-inits[[i]]$R[2]
        if(size==3){
          inits.internal[[i]][[15]] <-inits[[i]]$R[2]; 
          inits.internal[[i]][[16]] <-inits[[i]]$R[4]; 
          inits.internal[[i]][[17]] <-inits[[i]]$R[5]}
      }
      lower <- inits.internal[[i]][[15]]*inits.internal[[i]][[16]]-
        sqrt((1-inits.internal[[i]][[15]]^2)*(1-inits.internal[[i]][[16]]^2))
      upper <- inits.internal[[i]][[15]]*inits.internal[[i]][[16]]+
        sqrt((1-inits.internal[[i]][[15]]^2)*(1-inits.internal[[i]][[16]]^2))
      if(inits.internal[[i]][[17]]<lower | inits.internal[[i]][[17]]>upper)
        inits.internal[[i]][[17]] <- runif(1, lower, upper)
    }
  }}
  op<- system.file("bugs", "sep_2z.bug", package="zoib") 
  model <- jags.model(op, data=dataIn,n.adapt=0, inits=inits.internal, n.chains=n.chain)  
  return(model)
}