Parameter estimation of MCAR model with R

@Syukri_Adnan_Sangadj wrote:

hi kindness, I perform spatial data analysis with MCAR model. and so far I got some problem code to estimate the MCAR paremeter. below is my code using R

data.jatim=Data.Ekonometrik
n <- 38 ##number of locations
q <- 3 ##number of outcomes at each location
nltr <- q*(q+1)/2 ##number of triangular elements in the cross-covariance matrix
coords <- as.matrix(data.jatim[,c("Koor.X","Koor.Y")])
var.dependen<-as.matrix(data.jatim[c("Y1","Y2","Y3")])
cov(var.dependen)

n.samples=2000
model<-list(var.dependen~x1+x2+x3+x4)
model.MCAR<-spMvLM(model,coords=coords,data=data.jatim,
                   starting=list("phi"=rep(3/500,q),"A"=rep(0.05,nltr)),
                   tuning=list("phi"=rep(0.3,q),"A"=rep(0.001,nltr)),
                   prior=list("phi.unif"=list(rep(3/2500),rep(3/100,q)),
                    modified.pp=TRUE,"Psi.IG"=list(rep(2,q), rep(0.05,q))),
                   cov.model="exponential",
                   n.samples=n.samples,verbose=TRUE,n.report=1000)

result of the above code is

Error in spMvLM(model, coords = coords, data = data.jatim, starting = list(phi = rep(3/500,  : 
  error: a valid prior for K or A must be specified

Posts: 1

Participants: 1

Read full topic