inst/simulations/simu_weight_byEffect_bin.R

library(snow)
library(OPWeight)
library(OPWpaper)


cl <- makeCluster(10, type = "MPI")		# start zcluster


clusterExport(cl,"weight_by_delta")
clusterExport(cl,"weight_binary")
clusterExport(cl,"weight_byEffect_bin")


dat <- read.csv("ranksProb_byEffect_m10000.csv",h = TRUE)		# this prob. are saved in an excel file
filterEffectVec <- c(seq(0, 1, .2), 2, 3, 5, 8)
mVal = 10000
del = .000001
aVal = .05

clusterExport(cl,"dat")
clusterExport(cl,"filterEffectVec")
clusterExport(cl,"mVal")
clusterExport(cl,"del")
clusterExport(cl,"aVal")



weightByEffect_null.2  <- parSapply(cl,1:length(filterEffectVec),weight_byEffect_bin,alpha=aVal,null=.2,m=mVal,delInterval=del,filterEffectVec=filterEffectVec, datByNull=dat[,1:10])
weightByEffect_null.5  <- parSapply(cl,1:length(filterEffectVec),weight_byEffect_bin,alpha=aVal,null=.5,m=mVal,delInterval=del,filterEffectVec=filterEffectVec, datByNull=dat[,11:20])
weightByEffect_null.75 <- parSapply(cl,1:length(filterEffectVec),weight_byEffect_bin,alpha=aVal,null=.75,m=mVal,delInterval=del,filterEffectVec=filterEffectVec, datByNull=dat[,21:30])
weightByEffect_null.9  <- parSapply(cl,1:length(filterEffectVec),weight_byEffect_bin,alpha=aVal,null=.9,m=mVal,delInterval=del,filterEffectVec=filterEffectVec, datByNull=dat[,31:40])
weightByEffect_null.99 <- parSapply(cl,1:length(filterEffectVec),weight_byEffect_bin,alpha=aVal,null=.99,m=mVal,delInterval=del,filterEffectVec=filterEffectVec, datByNull=dat[,41:50])


stopCluster(cl)


# # save data
# #---------------
# datWeightByEffect <- data.frame(weightByEffect_null.2, weightByEffect_null.5,
#                                 weightByEffect_null.75, weightByEffect_null.9,
#                                 weightByEffect_null.99)
# colNames <- c(paste("null.2_ef",filterEffectVec ,sep=""),paste("null.5_ef",filterEffectVec ,sep=""),paste("null.75_ef",filterEffectVec ,sep=""),
#               paste("null.9_ef",filterEffectVec ,sep=""),paste("null.99_ef",filterEffectVec ,sep=""))
# colnames(datWeightByEffect) <- colNames
# write.csv(datWeightByEffect,file="weight_byEffect_bin_m10000.csv",row.names = F)
#
#
# dat = read.csv("weight_byEffect_bin_m10000.csv",h=T)
# weightByEffect = dat[,31:40]
#
# windows()
# m = 10000
# weightByEffect  <- weightByEffect_null.99
# apply(weightByEffect, 2, sum)
# par(mfrow = c(3, 3))
# for(i in 1:9)
# {
#     w <- weightByEffect[,i]
#     ww <- ifelse(rep(sum(w),m) == 0, 1, w/sum(w)*length(w))	# normalizing weight
#     plot(1:length(ww), ww, type="l")
#     print(sum(ww))
# }

#####################################################################################################################################

save.image("simu_weight_byEffect_bin.RData")
mshasan/OPWpaper documentation built on March 3, 2021, 7:02 a.m.