Hi all,
I am trying to use snow package to do a parallel MCMC. I have read a few
guides and articles, the following is that I came up with.
When I run it I got the error message:
Error in checkForRemoteErrors(lapply(cl, recvResult)) :
4 nodes produced errors; first error: could not find function ui.Next
The data is a longitudinal data with few repeated readings on a number of
individuals. However, the response is organised in a vector rather than a
matrix. E.g. (y_11 , y_12 , y_13 , y_14 , y_21 , y_22 , ... , y_n1 , ... ,
y_nTn )^T
X , Z are covariates.
beta is a matrix of coefficients associated with X
C is the latent class membership indicator
sigma.sq is the diagonal elements of the covariance matrix of u, which is a
random effect parameter and the one I am trying to sample with the function.
sd is the diagonal elements of the covariance matrix of the proposal
distribution.
The following is the particular function:
ui.Full.Sample.Cluster = function( levels , Y , X , beta , Z , C , sigma.sq
, sd , burnin , iteration )
{
cluster = makeCluster( 4 , type = MPI);
clusterSetupRNG( cluster );
patients = unique( levels );
q = length( X[1,] );
u = ui.Ini( q , length( patients ) );
n = levels[ length(patients) ];
marker = levels == patients[1];
y = Y[ marker ];
x = X[ marker, ];
z = Z[ marker, ];
u[1,] = ui.1.Sample( u[1,] , y , x , beta , z , C[1] , sigma.sq , sd ,
burnin , iteration )$uFinal;
for( i in 2:n)
{
marker = levels == patients[i];
y = Y[ marker ];
x = X[ marker, ];
z = Z[ marker, ];
print( i );
u[i, ] = clusterCall( cluster , ui.1.Sample, u[i,] , y , x ,
beta , z ,
C[i] , sigma.sq , sd , burnin , iteration )$uFinal;
print( i );
}
stopCluster( cluster );
u;
}
If anyone could help that would be much appreciated! But big thanks for
taking your time to read the post!
TelM8
--
View this message in context:
http://r.789695.n4.nabble.com/Problem-with-clusterCall-Error-in-checkForRemoteErrors-lapply-cl-recvResult-tp2338375p2338375.html
Sent from the R help mailing list archive at Nabble.com.
__
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
