rm(list=ls())

ClaimsData<-read.table("ClaimsYear1-4DIHandLCandSexandAge.csv",sep=",",header=TRUE)

# Order PCG by the largest average DIHY2 to smallest average DIHY2
pcgMat<-matrix(rep(0,16*45),ncol=16)
for (i in 1:45){
pcgMat[i,seq(1,16)]<-colSums(ClaimsData[ClaimsData[,35+i]==1,seq(116,131)])
}

MY2<-((pcgMat%*%seq(1,16))/rowSums(pcgMat))-1
rm(pcgMat)




# Order PCG by the largest average DIHY3 to smallest average DIHY3
pcgMat<-matrix(rep(0,16*45),ncol=16)
for (i in 1:45){
pcgMat[i,seq(1,16)]<-colSums(ClaimsData[ClaimsData[,35+i]==1,seq(132,147)])
}

MY3<-((pcgMat%*%seq(1,16))/rowSums(pcgMat))-1
rm(pcgMat)

M<-(MY2+MY3)/2

# orderedMY2<-cbind(order(MY2),M[order(MY2)])
# orderedMY3<-cbind(order(MY3),M[order(MY3)])

rm(MY2)
rm(MY3)
rm(i)

orderedM<-cbind(order(M),M[order(M)])






Claims<-read.table("Claims.csv",sep=",",header=TRUE)
Claims$PrimaryConditionGroup<-factor(Claims$PrimaryConditionGroup)
levels(Claims$PrimaryConditionGroup)<-seq(1,46)
Claims$PrimaryConditionGroup<-as.numeric(Claims$PrimaryConditionGroup)
Claims$PrimaryConditionGroup<-(Claims$PrimaryConditionGroup-1)
Claims$PrimaryConditionGroup[Claims$PrimaryConditionGroup==0]<-NA
Claims<-na.omit(Claims)

Claims$DIHPred<-rep(0,dim(Claims)[1])

for (i in 1:45){
Claims[Claims[,11]==i,15]<-M[i]
}

Claims<-Claims[order(Claims$MemberID),]