function to run MClust models in parallel using doMC

data_mining/parallelMclust.R

@@ -0,0 +1,36 @@
+#This was adapted from https://hpc-forge.cineca.it/files/CoursesDev/public/2015/Data_Analysis_School/script%20R/modelBased_genData.R
+#library(e1071)
+#library(parallel)
+library(foreach)
+library(doMC)
+library(mclust)
+
+parallelMcl <- function (
+                         data,
+                         ncores=4,
+                         G=2:6,
+                         modelNames=c("EII","EEI","EEE","VVV")
+                         ) {
+
+
+    param=expand.grid(G=G,modelNames=modelNames) 
+
+    parallel.function <- function(i) { Mclust( data=data, G=as.character(param$G[i]), modelNames=as.character(param$modelNames[i]),prior=priorControl()) }
+
+    registerDoMC(ncores)
+
+    b=Sys.time()
+    all <- foreach(i = 1:nrow(param)) %dopar% parallel.function(i)
+    e=Sys.time()
+    best = all[[which.max(sapply(all,function(x){x$bic}))]]
+    time=e-b
+    rm(e,b)
+    print(time)
+
+    print(ls(best))
+
+    entropy=sum((table(best$cl)/length(best$cl))*log(table(best$cl)/length(best$cl)))
+    print(entropy)
+    return(list(all=all, best=best, entropy=entropy, time=time))
+}
+