Nothing
R/Type.R
In cascsim: Casualty Actuarial Society Individual Claim Simulator
#' An S4 class to represent a claim type.
#'
#' @slot simno The simulation index.
#' @slot line A string to identify the business line that the claim belongs to.
#' @slot claimType A string to identify the type of the claim. It further classifies the claims within a business line. For example, the type could be based on the size of the loss.
#' @slot iRBNER A Boolean variable to indicate whether RBNER (open claims) should be simulated.
#' @slot iROPEN A Boolean variable to indicate whether claim reopen should be simulated.
#' @slot iIBNR A Boolean variable to indicate whether IBNR claims should be simulated.
#' @slot iUPR A Boolean variable to indicate whether future claims should be simulated.
#' @slot fIBNER IBNER development factor.
#' @slot severity Severity distribution.
#' @slot frequency Frequency distribution.
#' @slot reportLag Report lag distribution.
#' @slot settlementLag Settlement lag distribution.
#' @slot reopen Claim reopen probability based on the number of years after settlement till valuation date.
#' @slot reopenLag Reopen lag distribution.
#' @slot resettleLag Resettlement lag distribution.
#' @slot roDevFac Reopened claim development factor.
#' @slot ioDevFac A numeric variable to indicate the method of loss development for open claim severity. 1: Conditional distribution based on paid loss; 2: conditional distribution based on incurred loss; 3: year-to-year development factors
#' @slot irDevFac A numeric variable to indicate the method of loss development for claim reopen severity simulation. 1: Conditional distribution based on paid loss; 2: conditional distribution based on incurred loss; 3: year-to-year development factors
#' @slot freqIndex Frequency distribution time index.
#' @slot severityIndex Severity distribution time index.
#' @slot exposureIndex Exposure time index for IBNR or UPR.
#' @slot iCopula Whether copula is used to model severity, report lag and settlement lag.
#' @slot ssrCopula Copula object used for severity, report lag and settlement lag.
#' @slot sdata Indicating whether only closed claims (CLOSED) or closed + open claims (ALL) will be used for severity fitting.
#' @slot p0 An yearly table that controls the probability of invalid claim, excluding these valid claims less than deductible based on development year. It is based on the DevFac class.
setClass("ClaimType",
slots=c(
simno="character",
line="character",
claimType="character",
# claimClass="character",
iRBNER="logical",
iROPEN="logical",
iIBNR="logical",
iUPR="logical",
fIBNER="DevFac",
severity="Distribution",
frequency="Distribution",
reportLag="Distribution",
settlementLag="Distribution",
reopen="DevFac",
reopenLag="Distribution",
resettleLag="Distribution",
roDevFac="DevFac",
ioDevFac="numeric",
irDevFac="numeric",
IBNRfreqIndex="Index",
UPRfreqIndex="Index",
severityIndex="Index",
exposureIndex="Index",
iCopula="logical",
ssrCopula="CopulaObj",
laeDevFac="DevFac",
deductible="Distribution",
limit="Distribution",
sdata="character",
p0="DevFac"
),
prototype=list(
simno="XXX",
line="Line1",
claimType="Normal",
# claimClass="RBNER",
iRBNER=TRUE,
iROPEN=TRUE,
iIBNR=TRUE,
iUPR=TRUE,
fIBNER=new("DevFac",meanList=c(1.2,1.15,1.1,1.05,1),volList=c(runif(4)/10,0)),
severity=new("Lognormal",p1=3,p2=2),
frequency=new("Poisson",p1=100),
reportLag=new("Exponential",p1=0.25),
settlementLag=new("Exponential",p1=0.1),
reopen=new("DevFac",meanList=c(0.01,0.005,0.002,0.001,0),volList=rep(0,5)),
reopenLag=new("Exponential",p1=0.005),
resettleLag=new("Exponential",p1=0.05),
roDevFac=new("DevFac",meanList=c(1.05,1.1,1,1,1),volList=c(runif(4)/100,0)),
ioDevFac=3,
irDevFac=3,
IBNRfreqIndex=new("Index",monthlyIndex=rep(0,360)),
UPRfreqIndex=new("Index",monthlyIndex=rep(0,360)),
severityIndex=new("Index",monthlyIndex=rep(1,360)),
exposureIndex=new("Index",monthlyIndex=rep(1,360)),
iCopula=FALSE,
ssrCopula=new("CopulaObj",param=c(0,0,0),marginal=list(dist1=new("Lognormal",p1=3,p2=2),dist2=new("Exponential",p1=0.25),dist3=new("Exponential",p1=0.1)),dimension=3),
laeDevFac=new("DevFac",FacModel=TRUE,paras=c(0.1,0.01,0.005,0.002,5)),
deductible=new("Empirical",empirical=matrix(c(0,1,0,0),2,2)),
limit=new("Empirical",empirical=matrix(c(0,1,1e10,1e10),2,2)),
sdata="CLOSED",
p0=new("DevFac",meanList=c(0,0),volList=c(0,0))
)
)
#' Claim simulation at line/type/status level
#' @name claimSample
#' @param object ClaimType object
#' @param ... Additional parameters that may or may not be used.
#' @examples
#' #run time is about 12s(>10s) and is commented out here to avoid long waiting time
#' #library(cascsim)
#' #data(claimdata)
#' ##IBNR simulation
#' #claimobj <- new("ClaimType", line="Auto",claimType="N",iRBNER=FALSE,iROPEN=FALSE,
#' #iIBNR=TRUE,iUPR=FALSE,
#' #IBNRfreqIndex=new("Index",startDate=as.Date("2016-01-01"),
#' #monthlyIndex=rep(30,12)),iCopula=TRUE)
#' #ibnrdata <- claimSample(claimobj,claimdata)
#' #ibnrdata
#' @rdname claimSample-methods
#' @exportMethod claimSample
setGeneric("claimSample", function(object, ...) standardGeneric("claimSample"))
#' @param claimData claim data including existing claims for RBNER and claim reopenness analysis;
#' @param startDate Date from which claim data is included in the analysis;
#' @param evaluationDate Date of evaluation.
#' @rdname claimSample-methods
#' @aliases claimSample,ANY-method
setMethod("claimSample", signature("ClaimType"), function(object, claimData=data.frame(), startDate=as.Date("2012-01-01"), evaluationDate=as.Date("2016-12-31"))
{
tryCatch({
startDate <- toDate(startDate)
evaluationDate <- toDate(evaluationDate)
if (nrow(claimData)>0) {
claimData <- claimData[claimData[,"Type"] == object@claimType & claimData[,"LoB"] == object@line,]
claimData[,"status"] = toupper(claimData[,"status"])
claimData[,"claimLiability"] = toupper(claimData[,"claimLiability"])
claimData[,"occurrenceDate"] = as.character(toDate(claimData[,"occurrenceDate"]))
claimData[,"reportDate"] = as.character(toDate(claimData[,"reportDate"]))
claimData[,"settlementDate"] = as.character(toDate(claimData[,"settlementDate"]))
claimData <- claimData[as.Date(as.character(claimData[,"occurrenceDate"]))>=as.Date(as.character(startDate)),]
claimData <- claimData[as.Date(claimData[,"reportDate"])<=as.Date(evaluationDate),]
}
nid <- 0
sobj <- object@severity
#sobj@truncated <- FALSE
tmpdata = data.frame(ClaimID=character(),
LoB=character(),
Type=character(),
status=character(),
occurrenceDate=character(),
reportDate=character(),
incurredLoss=double(),
osRatio=double(),
settlementDate=character(),
totalLoss=double(),
ultimateLoss=double(),
Deductible=double(),
Limit=double(),
Paid=double(),
LAE=double(),
claimLiability=character(),
expectedLoss=double(),
ultimateLAE=double(),
expectedLAE=double(),
reopenDate=character(),
resettleDate=character(),
reopenLoss=double(),
Sim=double(),
stringsAsFactors=FALSE)
if (object@iRBNER == TRUE) {
if(nrow(claimData)==0){
nobs<-0
} else {
tmpdata <- claimData[claimData[,"status"] == "OPEN" | as.Date(claimData[,"settlementDate"])>as.Date(evaluationDate),]
nobs <- nrow(tmpdata)
}
if (nobs > 0){
occurrenceDates<-as.character(tmpdata[,"occurrenceDate"])
reportDates<-as.character(tmpdata[,"reportDate"])
osRatios<-as.numeric(tmpdata[,"osRatio"])
incurredLosses<-as.numeric(tmpdata[,"incurredLoss"])
settlementDates<-as.character(as.Date(reportDates) + as.numeric(lapply(setMin(object@settlementLag, as.numeric(as.Date(evaluationDate)-as.Date(reportDates))),doSample, n=1)) + as.numeric(as.Date(evaluationDate+1)-as.Date(reportDates))+1)
sindex<-getIndex(object@severityIndex,settlementDates)/getIndex(object@severityIndex,evaluationDate)
tmpdata[,"settlementDate"]<-settlementDates
developmentYears<-ceiling(as.numeric(as.Date(evaluationDate) - as.Date(tmpdata[,"occurrenceDate"]))/365)
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(settlementDates) - as.Date(tmpdata[,"occurrenceDate"]))/365))
if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "exponential" & object@ioDevFac==3){
eibnerdev<-object@fIBNER@paras[1]+object@fIBNER@paras[2]*developmentYears+object@fIBNER@paras[3]*incurredLosses+object@fIBNER@paras[4]*osRatios
ibnerdev<-eibnerdev+rnorm(nobs,0,object@fIBNER@paras[length(object@fIBNER@paras)])
if (length(object@fIBNER@xname)>0){
for (i in c(1:length(object@fIBNER@xname))){
ibnerdev<-ibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
eibnerdev<-eibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
}
}
ultimateLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(log(ibnerdev),0)
expectedLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(log(eibnerdev),0)
} else if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "log" & object@ioDevFac==3){
eibnerdev<-object@fIBNER@paras[1]+object@fIBNER@paras[2]*developmentYears+object@fIBNER@paras[3]*incurredLosses+object@fIBNER@paras[4]*osRatios
ibnerdev<-eibnerdev+rnorm(nobs,0,object@fIBNER@paras[length(object@fIBNER@paras)])
if (length(object@fIBNER@xname)>0){
for (i in c(1:length(object@fIBNER@xname))){
ibnerdev<-ibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
eibnerdev<-eibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
}
}
ultimateLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(exp(ibnerdev),0)
expectedLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(exp(eibnerdev),0)
} else if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "inverse" & object@ioDevFac==3){
eibnerdev<-object@fIBNER@paras[1]+object@fIBNER@paras[2]*developmentYears+object@fIBNER@paras[3]*incurredLosses+object@fIBNER@paras[4]*osRatios
ibnerdev<-eibnerdev+rnorm(nobs,0,object@fIBNER@paras[length(object@fIBNER@paras)])
if (length(object@fIBNER@xname)>0){
for (i in c(1:length(object@fIBNER@xname))){
ibnerdev<-ibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
eibnerdev<-eibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
}
}
ultimateLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(1/ibnerdev,0)
expectedLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(1/eibnerdev,0)
} else if (object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "identity" & object@ioDevFac==3) {
eibnerdev<-object@fIBNER@paras[1]+object@fIBNER@paras[2]*developmentYears+object@fIBNER@paras[3]*incurredLosses+object@fIBNER@paras[4]*osRatios
ibnerdev<-eibnerdev+rnorm(nobs,0,object@fIBNER@paras[length(object@fIBNER@paras)])
if (length(object@fIBNER@xname)>0){
for (i in c(1:length(object@fIBNER@xname))){
ibnerdev<-ibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
eibnerdev<-eibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
}
}
ultimateLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(ibnerdev,0)
expectedLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(eibnerdev,0)
} else if (object@ioDevFac==3){
years <- cbind(developmentYears, settlementYears)
ibnerdev<-ultiDevFac(years,meanDevFac=object@fIBNER@meanList,sdDevFac=object@fIBNER@volList,distType=object@fIBNER@distType)
eibnerdev<-ultiDevFac(years,meanDevFac=object@fIBNER@meanList,distType=object@fIBNER@distType)
ultimateLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(ibnerdev,0)
expectedLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(eibnerdev,0)
} else if (object@ioDevFac==2){
tmpdata[,"Limit"] <- ifelse(is.na(tmpdata[,"Limit"]),1e10,tmpdata[,"Limit"])
tmpdata[,"Deductible"] <- ifelse(is.na(tmpdata[,"Deductible"]),0,tmpdata[,"Deductible"])
ultimateLosses<-pmax(as.numeric(tmpdata[,"incurredLoss"]),pmax((pmin(tmpdata[,"Limit"]+tmpdata[,"Deductible"],doSample(sobj,nobs))-tmpdata[,"Deductible"]),0))
basicmean<-sampleMean(object@severity)
expectedLosses<-rep(basicmean,nobs)
} else {
tmpdata[,"Limit"] <- ifelse(is.na(tmpdata[,"Limit"]),1e10,tmpdata[,"Limit"])
tmpdata[,"Deductible"] <- ifelse(is.na(tmpdata[,"Deductible"]),0,tmpdata[,"Deductible"])
if(object@iCopula==FALSE){
probs <- Probability(sobj,tmpdata[,"Paid"])
newprobs <- probs + (1-probs)*runif(nobs)
ultimateLosses<-pmax(as.numeric(tmpdata[,"Paid"]),pmax((pmin(tmpdata[,"Limit"]+tmpdata[,"Deductible"],Quantile(sobj,newprobs))-tmpdata[,"Deductible"]),0))
} else {
for (j in c(1:nrow(tmpdata))) {
settlemin <- as.numeric(as.Date(evaluationDate)-as.Date(tmpdata[j,"reportDate"]))
sobjtrunc <- object@ssrCopula@marginal[[1]]@truncated
sobjmin <- object@ssrCopula@marginal[[1]]@min
robjtrunc <- object@ssrCopula@marginal[[2]]@truncated
robjmin <- object@ssrCopula@marginal[[2]]@min
object@ssrCopula@marginal[[1]]@min <- tmpdata[j,"Paid"]
object@ssrCopula@marginal[[1]]@truncated <- TRUE
object@ssrCopula@marginal[[2]]@min <- settlemin
object@ssrCopula@marginal[[2]]@truncated <- TRUE
samples <- copulaSample(object@ssrCopula,1)
tmpdata[j,"settlementDate"] <- as.character(as.Date(as.Date(tmpdata[j,"reportDate"]) + round(samples[,2])))
tmpdata[j,"ultimateLoss"] <- samples[,1]
}
object@ssrCopula@marginal[[1]]@min <- sobjmin
object@ssrCopula@marginal[[1]]@truncated <- sobjtrunc
object@ssrCopula@marginal[[2]]@min <- robjmin
object@ssrCopula@marginal[[2]]@truncated <- robjtrunc
settlementDates <- tmpdata[,"settlementDate"]
ultimateLosses <- tmpdata[,"ultimateLoss"]
}
basicmean<-sampleMean(object@severity)
expectedLosses<-rep(basicmean,nobs)
}
tmpdata[,"Limit"] <- ifelse(is.na(tmpdata[,"Limit"]),1e10,tmpdata[,"Limit"])
tmpdata[,"ultimateLoss"]<-pmin(round(ultimateLosses*sindex,2),tmpdata[,"Limit"])
zeros<-simP0(settlementYears,object@p0@meanList)
exzeros<-expectZeros(settlementYears,object@p0@meanList)
tmpdata[,"expectedLoss"]<-pmin(round(expectedLosses*sindex,2),tmpdata[,"Limit"])*(1-exzeros)
tmpdata[,"claimLiability"] <- ifelse(zeros==0,FALSE,TRUE)
tmpdata[,"ultimateLoss"]<- ifelse(tmpdata[,"claimLiability"] == FALSE, 0, tmpdata[,"ultimateLoss"])
if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "exponential" & object@ioDevFac==3){
laes<-as.numeric(tmpdata[,"LAE"])*pmax(log(ibnerdev)*sindex,0)
elaes<-as.numeric(tmpdata[,"LAE"])*pmax(log(eibnerdev)*sindex,0)
} else if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "log" & object@ioDevFac==3){
laes<-as.numeric(tmpdata[,"LAE"])*pmax(exp(ibnerdev)*sindex,0)
elaes<-as.numeric(tmpdata[,"LAE"])*pmax(exp(eibnerdev)*sindex,0)
} else if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "inverse" & object@ioDevFac==3){
laes<-as.numeric(tmpdata[,"LAE"])*pmax(1/ibnerdev*sindex,0)
elaes<-as.numeric(tmpdata[,"LAE"])*pmax(1/eibnerdev*sindex,0)
} else if (object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "identity" & object@ioDevFac==3) {
laes<-as.numeric(tmpdata[,"LAE"])*pmax(ibnerdev*sindex,0)
elaes<-as.numeric(tmpdata[,"LAE"])*pmax(eibnerdev*sindex,0)
} else if (object@ioDevFac==3){
laes<-as.numeric(tmpdata[,"LAE"])*pmax(ibnerdev*sindex,0)
elaes<-as.numeric(tmpdata[,"LAE"])*pmax(eibnerdev*sindex,0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "exponential" & object@ioDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(log(laedev),0)
elaes<-pmax(log(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "log" & object@ioDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(exp(laedev),0)
elaes<-pmax(exp(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "inverse" & object@ioDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(1/laedev,0)
elaes<-pmax(1/elaedev,0)
} else if (object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "identity" & object@ioDevFac<3) {
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(laedev,0)
elaes<-pmax(elaedev,0)
} else {
years <-cbind(developmentYears, settlementYears)
laedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList,sdDevFac=object@laeDevFac@volList, distType=object@laeDevFac@distType)
elaedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList, distType=object@laeDevFac@distType)
laes<-as.numeric(tmpdata[,"ultimateLoss"])*pmax(laedev,0)
elaes<-as.numeric(tmpdata[,"expectedLoss"])*pmax(elaedev,0)
}
tmpdata[,"ultimateLAE"]<-round(laes,2)#*sindex
tmpdata[,"expectedLAE"]<-round(elaes,2)#*sindex
tmpdata[,"reopenDate"]<-NA
tmpdata[,"resettleDate"]<-NA
tmpdata[,"reopenLoss"]<-NA
tmpdata[,"Sim"]<-object@simno
tmpdata<-tmpdata[,c("ClaimID","LoB","Type","status","occurrenceDate","reportDate","incurredLoss","osRatio","settlementDate","totalLoss","ultimateLoss","Deductible",
"Limit","Paid","LAE","claimLiability","expectedLoss","ultimateLAE","expectedLAE","reopenDate","resettleDate","reopenLoss","Sim")]
#return(claimData)
} else {
tmpdata = data.frame(ClaimID=character(),
LoB=character(),
Type=character(),
status=character(),
occurrenceDate=character(),
reportDate=character(),
incurredLoss=double(),
osRatio=double(),
settlementDate=character(),
totalLoss=double(),
ultimateLoss=double(),
Deductible=double(),
Limit=double(),
Paid=double(),
LAE=double(),
claimLiability=character(),
expectedLoss=double(),
ultimateLAE=double(),
expectedLAE=double(),
reopenDate=character(),
resettleDate=character(),
reopenLoss=double(),
Sim=double(),
stringsAsFactors=FALSE)
#return(noData)
}
}
if (object@iROPEN == TRUE) {
if(nrow(claimData)==0){
nobs<-0
} else {
claimData <- claimData[claimData[,"status"] == "CLOSED",]# | as.Date(claimData[,"settlementDate"])<=as.Date(evaluationDate),]
nobs <- nrow(claimData)
}
if(nobs>0){
settlementDates<-as.Date(claimData[,"settlementDate"])
occurrenceDates<-as.character(claimData[,"occurrenceDate"])
reportDates<-as.character(claimData[,"reportDate"])
osRatios<-as.numeric(claimData[,"osRatio"])
incurredLosses<-as.numeric(claimData[,"incurredLoss"])
closelags <- as.numeric(as.Date(evaluationDate) - as.Date(claimData[,"settlementDate"]))
closeyears <- pmax(1,ceiling(closelags/365))
developmentYears<-ceiling(as.numeric(as.Date(evaluationDate) - as.Date(claimData[,"occurrenceDate"]))/365)
if(object@reopen@FacModel==TRUE & object@reopen@fun == "exponential"){
reopenp<-object@reopen@paras[1]+object@reopen@paras[2]*developmentYears+object@reopen@paras[3]*incurredLosses+object@reopen@paras[4]*osRatios+rnorm(nobs,0,object@reopen@paras[length(object@reopen@paras)])
if (length(object@reopen@xname)>0){
for (i in c(1:length(object@reopen@xname))){
reopenp<-reopenp+object@reopen@paras[4+i]*as.numeric(claimData[,object@reopen@paras[i]])
}
}
reopens <- ifelse(runif(length(closeyears))<=log(reopenp),1,0)
reopenDates <- ifelse(reopens == 0, NA, as.character(settlementDates + as.numeric(lapply(setMin(object@reopenLag, as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))),doSample, n=1)) + as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))) + 1)
resettleDates <- ifelse(reopens == 0, NA, as.character(reopenDates + as.numeric(lapply(object@resettleLag,doSample, n=1))))
sindex<-getIndex(object@severityIndex,resettleDates)/getIndex(object@severityIndex,as.character(settlementDates))
} else if(object@reopen@FacModel==TRUE & object@reopen@fun == "log"){
reopenp<-object@reopen@paras[1]+object@reopen@paras[2]*developmentYears+object@reopen@paras[3]*incurredLosses+object@reopen@paras[4]*osRatios+rnorm(nobs,0,object@reopen@paras[length(object@reopen@paras)])
if (length(object@reopen@xname)>0){
for (i in c(1:length(object@reopen@xname))){
reopenp<-reopenp+object@reopen@paras[4+i]*as.numeric(claimData[,object@reopen@paras[i]])
}
}
reopens <- ifelse(runif(length(closeyears))<=exp(reopenp),1,0)
reopenDates <- ifelse(reopens == 0, NA, as.character(settlementDates + as.numeric(lapply(setMin(object@reopenLag, as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))),doSample, n=1)) + as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))) + 1)
resettleDates <- ifelse(reopens == 0, NA, as.character(reopenDates + as.numeric(lapply(object@resettleLag,doSample, n=1))))
sindex<-getIndex(object@severityIndex,resettleDates)/getIndex(object@severityIndex,as.character(settlementDates))
} else if(object@reopen@FacModel==TRUE & object@reopen@fun == "inverse"){
reopenp<-object@reopen@paras[1]+object@reopen@paras[2]*developmentYears+object@reopen@paras[3]*incurredLosses+object@reopen@paras[4]*osRatios+rnorm(nobs,0,object@reopen@paras[length(object@reopen@paras)])
if (length(object@reopen@xname)>0){
for (i in c(1:length(object@reopen@xname))){
reopenp<-reopenp+object@reopen@paras[4+i]*as.numeric(claimData[,object@reopen@paras[i]])
}
}
reopens <- ifelse(runif(length(closeyears))<=1/reopenp,1,0)
reopenDates <- ifelse(reopens == 0, NA, as.character(settlementDates + as.numeric(lapply(setMin(object@reopenLag, as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))),doSample, n=1)) + as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))) + 1)
resettleDates <- ifelse(reopens == 0, NA, as.character(reopenDates + as.numeric(lapply(object@resettleLag,doSample, n=1))))
sindex<-getIndex(object@severityIndex,resettleDates)/getIndex(object@severityIndex,as.character(settlementDates))
} else if (object@reopen@FacModel==TRUE & object@reopen@fun == "identity") {
reopenp<-object@reopen@paras[1]+object@reopen@paras[2]*developmentYears+object@reopen@paras[3]*incurredLosses+object@reopen@paras[4]*osRatios+rnorm(nobs,0,object@reopen@paras[length(object@reopen@paras)])
if (length(object@reopen@xname)>0){
for (i in c(1:length(object@reopen@xname))){
reopenp<-reopenp+object@reopen@paras[4+i]*as.numeric(claimData[,object@reopen@paras[i]])
}
}
reopens <- ifelse(runif(length(closeyears))<=reopenp,1,0)
reopenDates <- ifelse(reopens == 0, NA, as.character(settlementDates + as.numeric(lapply(setMin(object@reopenLag, as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))),doSample, n=1))+as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))+1))
resettleDates <- ifelse(reopens == 0, NA, as.character(reopenDates + as.numeric(lapply(object@resettleLag,doSample, n=1))))
sindex<-getIndex(object@severityIndex,resettleDates)/getIndex(object@severityIndex,as.character(settlementDates))
} else {
reopens<-rreopen(closeyears,object@reopen@meanList)
reopenDates <- ifelse(reopens == 0, NA, as.character(settlementDates + as.numeric(lapply(setMin(object@reopenLag, as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))),doSample, n=1))+as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))+1))
resettleDates <- ifelse(reopens == 0, NA, as.character(as.Date(reopenDates) + as.numeric(doSample(object@resettleLag, n=length(reopens)))))
sindex<-getIndex(object@severityIndex,resettleDates)/getIndex(object@severityIndex,as.character(settlementDates))
}
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(resettleDates) - as.Date(claimData[,"occurrenceDate"]))/365))
if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "exponential" & object@irDevFac==3){
erodev<-object@roDevFac@paras[1]+object@roDevFac@paras[2]*developmentYears+object@roDevFac@paras[3]*incurredLosses+object@roDevFac@paras[4]*osRatios
rodev<-erodev+rnorm(nobs,0,object@roDevFac@paras[length(object@roDevFac@paras)])
if (length(object@roDevFac@xname)>0){
for (i in c(1:length(object@roDevFac@xname))){
rodev<-rodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
erodev<-erodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
}
}
ultimateLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(log(rodev),0)
expectedLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(log(erodev),0)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses*sindex)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses*sindex)
} else if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "log" & object@irDevFac==3){
erodev<-object@roDevFac@paras[1]+object@roDevFac@paras[2]*developmentYears+object@roDevFac@paras[3]*incurredLosses+object@roDevFac@paras[4]*osRatios
rodev<-erodev+rnorm(nobs,0,object@roDevFac@paras[length(object@roDevFac@paras)])
if (length(object@roDevFac@xname)>0){
for (i in c(1:length(object@roDevFac@xname))){
rodev<-rodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
erodev<-erodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
}
}
ultimateLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(exp(rodev),0)
expectedLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(exp(erodev),0)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses*sindex)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses*sindex)
} else if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "inverse" & object@irDevFac==3){
erodev<-object@roDevFac@paras[1]+object@roDevFac@paras[2]*developmentYears+object@roDevFac@paras[3]*incurredLosses+object@roDevFac@paras[4]*osRatios
rodev<-erodev+rnorm(nobs,0,object@roDevFac@paras[length(object@roDevFac@paras)])
if (length(object@roDevFac@xname)>0){
for (i in c(1:length(object@roDevFac@xname))){
rodev<-rodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
erodev<-erodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
}
}
ultimateLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(1/rodev,0)
expectedLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(1/erodev,0)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses*sindex)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses*sindex)
} else if (object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "identity" & object@irDevFac==3) {
erodev<-object@roDevFac@paras[1]+object@roDevFac@paras[2]*developmentYears+object@roDevFac@paras[3]*incurredLosses+object@roDevFac@paras[4]*osRatios
rodev<-erodev+rnorm(nobs,0,object@roDevFac@paras[length(object@roDevFac@paras)])
if (length(object@roDevFac@xname)>0){
for (i in c(1:length(object@roDevFac@xname))){
rodev<-rodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
erodev<-erodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
}
}
ultimateLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(rodev,0)
expectedLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(erodev,0)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses*sindex)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses*sindex)
} else if (object@irDevFac==3){
years <- cbind(developmentYears, settlementYears)
rodev<-ultiDevFac(years,meanDevFac=object@roDevFac@meanList,sdDevFac=object@roDevFac@volList,distType=object@roDevFac@distType)
erodev<-ultiDevFac(years,meanDevFac=object@roDevFac@meanList,distType=object@roDevFac@distType)
ultimateLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(rodev,0)
expectedLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(erodev,0)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses*sindex)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses*sindex)
} else if (object@irDevFac==2){
claimData[,"Limit"] <- ifelse(is.na(claimData[,"Limit"]),1e10,claimData[,"Limit"])
claimData[,"Deductible"] <- ifelse(is.na(claimData[,"Deductible"]),0,claimData[,"Deductible"])
ultimateLosses<-pmax(as.numeric(claimData[,"incurredLoss"]),pmax((pmin(claimData[,"Limit"]+claimData[,"Deductible"],doSample(sobj,nobs)*sindex)-claimData[,"Deductible"]),0))
basicmean<-sampleMean(object@severity)
expectedLosses<-rep(basicmean,nobs)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses)
} else {
claimData[,"Limit"] <- ifelse(is.na(claimData[,"Limit"]),1e10,claimData[,"Limit"])
claimData[,"Deductible"] <- ifelse(is.na(claimData[,"Deductible"]),0,claimData[,"Deductible"])
probs <- Probability(sobj,claimData[,"Paid"])
newprobs <- probs + (1-probs)*runif(nobs)
ultimateLosses<-pmax(as.numeric(claimData[,"Paid"]),pmax((pmin(claimData[,"Limit"]+claimData[,"Deductible"],Quantile(sobj,newprobs)*sindex)-claimData[,"Deductible"]),0))
basicmean<-sampleMean(object@severity)
expectedLosses<-rep(basicmean,nobs)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses)
}
if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "exponential" & object@irDevFac==3){
laes<-as.numeric(claimData[,"LAE"])*pmax(log(rodev),0)
elaes<-as.numeric(claimData[,"LAE"])*pmax(log(erodev),0)
laes<- ifelse(reopens == 0, NA, laes*sindex)
elaes<- ifelse(reopens == 0, NA, elaes*sindex)
} else if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "log" & object@irDevFac==3){
laes<-as.numeric(claimData[,"LAE"])*pmax(exp(rodev),0)
elaes<-as.numeric(claimData[,"LAE"])*pmax(exp(erodev),0)
laes<- ifelse(reopens == 0, NA, laes*sindex)
elaes<- ifelse(reopens == 0, NA, elaes*sindex)
} else if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "inverse" & object@irDevFac==3){
laes<-as.numeric(claimData[,"LAE"])*pmax(1/rodev,0)
elaes<-as.numeric(claimData[,"LAE"])*pmax(1/erodev,0)
laes<- ifelse(reopens == 0, NA, laes*sindex)
elaes<- ifelse(reopens == 0, NA, elaes*sindex)
} else if (object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "identity" & object@irDevFac==3) {
laes<-as.numeric(claimData[,"LAE"])*pmax(rodev,0)
elaes<-as.numeric(claimData[,"LAE"])*pmax(erodev,0)
laes<- ifelse(reopens == 0, NA, laes*sindex)
elaes<- ifelse(reopens == 0, NA, elaes*sindex)
} else if (object@irDevFac==3){
laes<-as.numeric(claimData[,"LAE"])*pmax(rodev,0)
elaes<-as.numeric(claimData[,"LAE"])*pmax(erodev,0)
laes<- ifelse(reopens == 0, NA, laes*sindex)
elaes<- ifelse(reopens == 0, NA, elaes*sindex)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "exponential" & object@irDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(log(laedev),0)
elaes<-pmax(log(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "log" & object@irDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(exp(laedev),0)
elaes<-pmax(exp(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "inverse" & object@irDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(1/laedev,0)
elaes<-pmax(1/elaedev,0)
} else if (object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "identity") {
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(laedev,0)
elaes<-pmax(elaedev,0)
} else {
years <- cbind(developmentYears, settlementYears)
laedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList,sdDevFac=object@laeDevFac@volList, distType=object@laeDevFac@distType)
elaedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList, distType=object@laeDevFac@distType)
laes<-ultimateLosses*pmax(laedev,0)
elaes<-expectedLosses*pmax(elaedev,0)
}
claimData[,"ultimateLoss"]<-claimData[,"incurredLoss"]
claimData[,"expectedLoss"]<-round(expectedLosses,2)
claimData[,"ultimateLAE"]<-round(laes,2)#*sindex
claimData[,"expectedLAE"]<-round(elaes,2)#*sindex
claimData[,"reopenDate"]<-reopenDates
claimData[,"resettleDate"]<-resettleDates
claimData[,"Limit"] <- ifelse(is.na(claimData[,"Limit"]),1e10,claimData[,"Limit"])
claimData[,"reopenLoss"]<-pmin(round(ultimateLosses,2),claimData[,"Limit"])
claimData[,"Sim"]<-object@simno
claimData<-claimData[,c("ClaimID","LoB","Type","status","occurrenceDate","reportDate","incurredLoss","osRatio","settlementDate","totalLoss","ultimateLoss","Deductible",
"Limit","Paid","LAE","claimLiability","expectedLoss","ultimateLAE","expectedLAE","reopenDate","resettleDate","reopenLoss","Sim")]
tmpdata <- rbind(tmpdata,claimData)
}
}
if (object@iIBNR==TRUE) {
basicmean<-sampleMean(object@severity)
days <- c(31,28,31,30,31,30,31,31,30,31,30,31)
nid <- length(object@IBNRfreqIndex@monthlyIndex)
startyear <- substr(as.character(object@IBNRfreqIndex@startDate),1,4)
startmonth <- substr(as.character(object@IBNRfreqIndex@startDate),6,7)
ibnrdata = data.frame(ClaimID=character(),
LoB=character(),
Type=character(),
status=character(),
occurrenceDate=character(),
reportDate=character(),
incurredLoss=double(),
osRatio=double(),
settlementDate=character(),
totalLoss=double(),
ultimateLoss=double(),
Deductible=double(),
Limit=double(),
Paid=double(),
LAE=double(),
claimLiability=character(),
expectedLoss=double(),
ultimateLAE=double(),
expectedLAE=double(),
reopenDate=character(),
resettleDate=character(),
reopenLoss=double(),
Sim=double(),
stringsAsFactors=FALSE)
if(object@iCopula==TRUE){
ssindex<-vector()
for(i in c(1:nid)){
if(object@IBNRfreqIndex@monthlyIndex[i] > 0) {
currentmonth <- (i-1) %% 12 + as.numeric(startmonth)
currentyear <- floor((i-1)/12)+as.numeric(startyear)
if(currentmonth>12) {
currentmonth <- currentmonth-12
currentyear <- currentyear+1
}
umax <- days[currentmonth]
if(currentyear <10) {
currentmonth <- paste0("0",currentmonth)
}
occurrenceDate<-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01"))+round(runif(object@IBNRfreqIndex@monthlyIndex[i],min=0,max=(umax-1)))
ibnrs<-occurrenceDate
sobjtrunc <- object@ssrCopula@marginal[[1]]@truncated
robjtrunc <- object@ssrCopula@marginal[[3]]@truncated
robjmin <- object@ssrCopula@marginal[[3]]@min
object@ssrCopula@marginal[[1]]@truncated <- FALSE
object@ssrCopula@marginal[[3]]@min <- as.numeric(as.Date(evaluationDate)-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01")))
object@ssrCopula@marginal[[3]]@truncated <- TRUE
samples <- copulaSample(object@ssrCopula,object@IBNRfreqIndex@monthlyIndex[i])
object@ssrCopula@marginal[[1]]@truncated <- sobjtrunc
object@ssrCopula@marginal[[3]]@truncated <- robjtrunc
object@ssrCopula@marginal[[3]]@min <- robjmin
reportDate<-as.Date(occurrenceDate + round(samples[,3]) + as.numeric(as.Date(evaluationDate)-as.Date(occurrenceDate))+1)
settlementDate<-as.Date(reportDate + round(samples[,2]))
sindex<-getIndex(object@severityIndex,settlementDate)/getIndex(object@severityIndex,evaluationDate)
ssindex<-c(ssindex,sindex)
deductibles<-round(doSample(object@deductible,nrow(samples)),-2)
limits<-round(doSample(object@limit,nrow(samples)),-2)
totalLoss<-samples[,1]*sindex
ultimateLoss<-pmax(0,pmin(limits+deductibles,totalLoss)-deductibles)
expectedLoss<-basicmean*sindex
ibnrs<-cbind(as.character(occurrenceDate),as.character(reportDate),as.character(settlementDate),round(ultimateLoss,2),round(expectedLoss,2))
ibnrs<-as.data.frame(ibnrs)
colnames(ibnrs)<-c("occurrenceDate","reportDate","settlementDate","ultimateLoss","expectedLoss")
ibnrs$totalLoss<-totalLoss
ibnrs$ultimateLoss<-as.numeric(as.character(ibnrs$ultimateLoss))
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(settlementDate) - as.Date(occurrenceDate))/365))
zeros<-simP0(settlementYears,object@p0@meanList)
exzeros<-expectZeros(settlementYears,object@p0@meanList)
ibnrs$claimLiability <- ifelse(zeros==0,FALSE,TRUE)
ibnrs$ultimateLoss <- ifelse(ibnrs$claimLiability == FALSE, 0, ibnrs$ultimateLoss)
ibnrs[,"Deductible"]<-deductibles
ibnrs[,"Limit"]<-limits
ibnrs[,"Paid"]<-ibnrs$ultimateLoss
ibnrs$expectedLoss<-as.numeric(as.character(ibnrs$expectedLoss))*(1-exzeros)
ibnrs[,"LoB"]<-object@line
ibnrs[,"Type"]<-object@claimType
ibnrs[,"status"]<-"IBNR"
ibnrs[,"incurredLoss"]<-ibnrs$ultimateLoss
ibnrs[,"osRatio"]<-0
ibnrs[,"ClaimID"]<-as.character(paste0(i,"X",c(1:object@IBNRfreqIndex@monthlyIndex[i])))
ibnrdata<-rbind(ibnrdata,ibnrs)
}
}
} else {
ssindex<-vector()
for(i in c(1:nid)){
if(object@IBNRfreqIndex@monthlyIndex[i] > 0) {
currentmonth <- (i-1) %% 12 + as.numeric(startmonth)
currentyear <- floor((i-1)/12)+as.numeric(startyear)
if(currentmonth>12) {
currentmonth <- currentmonth-12
currentyear <- currentyear+1
}
umax <- days[currentmonth]
if(currentyear <10) {
currentmonth <- paste0("0",currentmonth)
}
occurrenceDate<-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01"))+round(runif(object@IBNRfreqIndex@monthlyIndex[i],min=0,max=(umax-1)))
ibnrs<-occurrenceDate
object@reportLag@min <- as.numeric(as.Date(evaluationDate)-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01")))
object@reportLag@truncated <- TRUE
reportDate<-as.Date(occurrenceDate + round(doSample(object@reportLag,object@IBNRfreqIndex@monthlyIndex[i])) + as.numeric(as.Date(evaluationDate)-as.Date(occurrenceDate))+1)
settlementDate<-as.Date(reportDate + round(doSample(object@settlementLag,object@IBNRfreqIndex@monthlyIndex[i])))
sindex<-getIndex(object@severityIndex,settlementDate)#/getIndex(object@severityIndex,evaluationDate)
ssindex<-c(ssindex,sindex)
deductibles<-round(doSample(object@deductible,object@IBNRfreqIndex@monthlyIndex[i]),-2)
limits<-round(doSample(object@limit,object@IBNRfreqIndex@monthlyIndex[i]),-2)
totalLoss<-doSample(sobj,object@IBNRfreqIndex@monthlyIndex[i])*sindex
ultimateLoss<-pmax(0,pmin(limits+deductibles,totalLoss)-deductibles)
expectedLoss<-basicmean*sindex
ibnrs<-cbind(as.character(occurrenceDate),as.character(reportDate),as.character(settlementDate),round(ultimateLoss,2),round(expectedLoss,2))
ibnrs<-as.data.frame(ibnrs)
colnames(ibnrs)<-c("occurrenceDate","reportDate","settlementDate","ultimateLoss","expectedLoss")
ibnrs$totalLoss<-totalLoss
ibnrs$ultimateLoss<-as.numeric(as.character(ibnrs$ultimateLoss))
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(settlementDate) - as.Date(occurrenceDate))/365))
zeros<-simP0(settlementYears,object@p0@meanList)
exzeros<-expectZeros(settlementYears,object@p0@meanList)
ibnrs$claimLiability <- ifelse(zeros==0,FALSE,TRUE)
ibnrs$ultimateLoss <- ifelse(ibnrs$claimLiability == FALSE, 0, ibnrs$ultimateLoss)
ibnrs[,"Deductible"]<-deductibles
ibnrs[,"Limit"]<-limits
ibnrs[,"Paid"]<-ibnrs$ultimateLoss
ibnrs$expectedLoss<-as.numeric(as.character(ibnrs$expectedLoss))*(1-exzeros)
ibnrs[,"LoB"]<-object@line
ibnrs[,"Type"]<-object@claimType
ibnrs[,"status"]<-"IBNR"
ibnrs[,"incurredLoss"]<-ibnrs$ultimateLoss
ibnrs[,"osRatio"]<-0
ibnrs[,"ClaimID"]<-as.character(paste0(i,"X",c(1:object@IBNRfreqIndex@monthlyIndex[i])))
ibnrdata<-rbind(ibnrdata,ibnrs)
}
}
}
if (nrow(ibnrdata)>0) {
developmentYears<-ceiling(as.numeric(as.Date(ibnrdata[,"reportDate"]) - as.Date(ibnrdata[,"occurrenceDate"]))/365)
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(ibnrdata[,"settlementDate"]) - as.Date(ibnrdata[,"occurrenceDate"]))/365))
incurredLosses<-ibnrdata[,"ultimateLoss"]
osRatios<-ibnrdata[,"osRatio"]
if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "exponential"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(ibnrdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(log(laedev),0)
elaes<-pmax(log(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "log"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(ibnrdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(exp(laedev),0)
elaes<-pmax(exp(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "inverse"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(ibnrdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(1/laedev,0)
elaes<-pmax(1/elaedev,0)
} else if (object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "identity") {
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(ibnrdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(laedev,0)
elaes<-pmax(elaedev,0)
} else {
years <- cbind(developmentYears,settlementYears)
laedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList,sdDevFac=object@laeDevFac@volList, distType=object@laeDevFac@distType)
elaedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList, distType=object@laeDevFac@distType)
laes<-as.numeric(ibnrdata[,"ultimateLoss"])*pmax(laedev,0)
elaes<-as.numeric(ibnrdata[,"expectedLoss"])*pmax(elaedev,0)
}
ibnrdata[,"ultimateLAE"]<-round(laes,2)#*ssindex
ibnrdata[,"expectedLAE"]<-round(elaes,2)#*ssindex
ibnrdata[,"reopenDate"]<-NA
ibnrdata[,"resettleDate"]<-NA
ibnrdata[,"reopenLoss"]<-NA
ibnrdata[,"LAE"]<-ibnrdata[,"ultimateLAE"]*1
ibnrdata[,"Sim"]<-object@simno
# ibnrdata<-ibnrdata[,c(11,6,7,8,1,2,9,10,3,4,17,5,12:16,18)]
ibnrdata<-ibnrdata[,c("ClaimID","LoB","Type","status","occurrenceDate","reportDate","incurredLoss","osRatio","settlementDate","totalLoss","ultimateLoss","Deductible",
"Limit","Paid","LAE","claimLiability","expectedLoss","ultimateLAE","expectedLAE","reopenDate","resettleDate","reopenLoss","Sim")]
tmpdata <- rbind(tmpdata,ibnrdata)
}
}
if (object@iUPR == TRUE) {
basicmean<-sampleMean(object@severity)
days <- c(31,28,31,30,31,30,31,31,30,31,30,31)
nid <- length(object@UPRfreqIndex@monthlyIndex)
startyear <- substr(as.character(object@UPRfreqIndex@startDate),1,4)
startmonth <- substr(as.character(object@UPRfreqIndex@startDate),6,7)
uprdata = data.frame(ClaimID=character(),
LoB=character(),
Type=character(),
status=character(),
occurrenceDate=character(),
reportDate=character(),
incurredLoss=double(),
osRatio=double(),
settlementDate=character(),
totalLoss=double(),
ultimateLoss=double(),
Deductible=double(),
Limit=double(),
Paid=double(),
LAE=double(),
claimLiability=character(),
expectedLoss=double(),
ultimateLAE=double(),
expectedLAE=double(),
reopenDate=character(),
resettleDate=character(),
reopenLoss=double(),
Sim=double(),
stringsAsFactors=FALSE)
if(object@iCopula==TRUE){
ssindex<-vector()
for(i in c(1:nid)){
if(object@UPRfreqIndex@monthlyIndex[i] > 0) {
currentmonth <- (i-1) %% 12 + as.numeric(startmonth)
currentyear <- floor((i-1)/12)+as.numeric(startyear)
if(currentmonth>12) {
currentmonth <- currentmonth-12
currentyear <- currentyear+1
}
umax <- days[currentmonth]
if(currentyear <10) {
currentmonth <- paste0("0",currentmonth)
}
occurrenceDate<-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01"))+round(runif(object@UPRfreqIndex@monthlyIndex[i],min=0,max=(umax-1)))
uprs<-occurrenceDate
sobjtrunc <- object@ssrCopula@marginal[[1]]@truncated
object@ssrCopula@marginal[[1]]@truncated <- FALSE
samples <- copulaSample(object@ssrCopula,object@UPRfreqIndex@monthlyIndex[i])
object@ssrCopula@marginal[[1]]@truncated <- sobjtrunc
reportDate<-as.Date(occurrenceDate + round(samples[,3]))# + as.numeric(as.Date(evaluationDate)-as.Date(occurrenceDate))+1)
settlementDate<-as.Date(reportDate + round(samples[,2]))
sindex<-getIndex(object@severityIndex,settlementDate)#/getIndex(object@severityIndex,evaluationDate)
ssindex<-c(ssindex,sindex)
deductibles<-round(doSample(object@deductible,nrow(samples)),-2)
limits<-round(doSample(object@limit,nrow(samples)),-2)
totalLoss<-samples[,1]*sindex
ultimateLoss<-pmax(0,pmin(limits+deductibles,totalLoss)-deductibles)
expectedLoss<-basicmean*sindex
uprs<-cbind(as.character(occurrenceDate),as.character(reportDate),as.character(settlementDate),round(ultimateLoss,2),round(expectedLoss,2))
uprs<-as.data.frame(uprs)
colnames(uprs)<-c("occurrenceDate","reportDate","settlementDate","ultimateLoss","expectedLoss")
uprs$totalLoss<-totalLoss
uprs$ultimateLoss<-as.numeric(as.character(uprs$ultimateLoss))
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(settlementDate) - as.Date(occurrenceDate))/365))
zeros<-simP0(settlementYears,object@p0@meanList)
exzeros<-expectZeros(settlementYears,object@p0@meanList)
uprs$claimLiability <- ifelse(zeros==0,FALSE,TRUE)
uprs$ultimateLoss <- ifelse(uprs$claimLiability == FALSE, 0, uprs$ultimateLoss)
uprs[,"Deductible"]<-deductibles
uprs[,"Limit"]<-limits
uprs[,"Paid"]<- uprs$ultimateLoss
uprs$expectedLoss<-as.numeric(as.character(uprs$expectedLoss))*(1-exzeros)
uprs[,"LoB"]<-object@line
uprs[,"Type"]<-object@claimType
uprs[,"status"]<-"UPR"
uprs[,"incurredLoss"]<-uprs$ultimateLoss
uprs[,"osRatio"]<-0
uprs[,"ClaimID"]<-as.character(paste0(i,"X",c(1:object@UPRfreqIndex@monthlyIndex[i])))
uprdata<-rbind(uprdata,uprs)
}
}
} else {
ssindex<-vector()
for(i in c(1:nid)){
if(object@UPRfreqIndex@monthlyIndex[i] > 0) {
currentmonth <- (i-1) %% 12 + as.numeric(startmonth)
currentyear <- floor((i-1)/12)+as.numeric(startyear)
if(currentmonth>12) {
currentmonth <- currentmonth-12
currentyear <- currentyear+1
}
umax <- days[currentmonth]
if(currentyear <10) {
currentmonth <- paste0("0",currentmonth)
}
occurrenceDate<-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01"))+round(runif(object@UPRfreqIndex@monthlyIndex[i],min=0,max=(umax-1)))
uprs<-occurrenceDate
# object@reportLag@min <- as.numeric(as.Date(evaluationDate)-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01")))
# object@reportLag@truncated <- TRUE
reportDate<-as.Date(occurrenceDate + round(doSample(object@reportLag,object@UPRfreqIndex@monthlyIndex[i])))
settlementDate<-as.Date(reportDate + round(doSample(object@settlementLag,object@UPRfreqIndex@monthlyIndex[i])))
sindex<-getIndex(object@severityIndex,settlementDate)/getIndex(object@severityIndex,evaluationDate)
ssindex<-c(ssindex,sindex)
deductibles<-round(doSample(object@deductible,object@UPRfreqIndex@monthlyIndex[i]),-2)
limits<-round(doSample(object@limit,object@UPRfreqIndex@monthlyIndex[i]),-2)
totalLoss<-doSample(sobj,object@UPRfreqIndex@monthlyIndex[i])*sindex
ultimateLoss<-pmax(0,pmin(limits+deductibles,totalLoss)-deductibles)
expectedLoss<-basicmean*sindex
uprs<-cbind(as.character(occurrenceDate),as.character(reportDate),as.character(settlementDate),round(ultimateLoss,2),round(expectedLoss,2))
uprs<-as.data.frame(uprs)
colnames(uprs)<-c("occurrenceDate","reportDate","settlementDate","ultimateLoss","expectedLoss")
uprs$totalLoss<-totalLoss
uprs$ultimateLoss<-as.numeric(as.character(uprs$ultimateLoss))
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(settlementDate) - as.Date(occurrenceDate))/365))
zeros<-simP0(settlementYears,object@p0@meanList)
exzeros<-expectZeros(settlementYears,object@p0@meanList)
uprs$claimLiability <- ifelse(zeros==0,FALSE,TRUE)
uprs$ultimateLoss <- ifelse(uprs$claimLiability == FALSE, 0, uprs$ultimateLoss)
uprs[,"Deductible"]<-deductibles
uprs[,"Limit"]<-limits
uprs[,"Paid"]<-uprs$ultimateLoss
uprs$expectedLoss<-as.numeric(as.character(uprs$expectedLoss))*(1-exzeros)
uprs[,"LoB"]<-object@line
uprs[,"Type"]<-object@claimType
uprs[,"status"]<-"UPR"
uprs[,"incurredLoss"]<-uprs$ultimateLoss
uprs[,"osRatio"]<-0
uprs[,"ClaimID"]<-as.character(paste0(i,"X",c(1:object@UPRfreqIndex@monthlyIndex[i])))
uprdata<-rbind(uprdata,uprs)
}
}
}
if (nrow(uprdata)>0) {
developmentYears<-ceiling(as.numeric(as.Date(uprdata[,"reportDate"]) - as.Date(uprdata[,"occurrenceDate"]))/365)
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(uprdata[,"settlementDate"]) - as.Date(uprdata[,"occurrenceDate"]))/365))
incurredLosses<-uprdata[,"ultimateLoss"]
osRatios<-uprdata[,"osRatio"]
if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "exponential"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(uprdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(log(laedev),0)
elaes<-pmax(log(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "log"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(uprdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(exp(laedev),0)
elaes<-pmax(exp(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "inverse"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(uprdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(1/laedev,0)
elaes<-pmax(1/elaedev,0)
} else if (object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "identity") {
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(uprdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(laedev,0)
elaes<-pmax(elaedev,0)
} else {
years <- cbind(developmentYears,settlementYears)
laedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList,sdDevFac=object@laeDevFac@volList, distType=object@laeDevFac@distType)
elaedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList, distType=object@laeDevFac@distType)
laes<-as.numeric(uprdata[,"ultimateLoss"])*pmax(laedev,0)
elaes<-as.numeric(uprdata[,"expectedLoss"])*pmax(elaedev,0)
}
uprdata[,"ultimateLAE"]<-round(laes,2)#*ssindex
uprdata[,"expectedLAE"]<-round(elaes,2)#*ssindex
uprdata[,"reopenDate"]<-NA
uprdata[,"resettleDate"]<-NA
uprdata[,"reopenLoss"]<-NA
uprdata[,"LAE"]<-uprdata[,"ultimateLAE"]*1
uprdata[,"Sim"]<-object@simno
uprdata<-uprdata[,c("ClaimID","LoB","Type","status","occurrenceDate","reportDate","incurredLoss","osRatio","settlementDate","totalLoss","ultimateLoss","Deductible",
"Limit","Paid","LAE","claimLiability","expectedLoss","ultimateLAE","expectedLAE","reopenDate","resettleDate","reopenLoss","Sim")]
tmpdata <- rbind(tmpdata,uprdata)
}
}
return(tmpdata)
}, error = function(err){
message(paste0(">>>Critical Error for claim simulation: Line-", object@line, " Type-", object@claimType, " Sim No.-", object@simno, ", ",err))
gc()
})
})
Try the cascsim package in your browser
Any scripts or data that you put into this service are public.
cascsim documentation built on Jan. 13, 2020, 5:07 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
#' An S4 class to represent a claim type.
#'
#' @slot simno The simulation index.
#' @slot line A string to identify the business line that the claim belongs to.
#' @slot claimType A string to identify the type of the claim. It further classifies the claims within a business line. For example, the type could be based on the size of the loss.
#' @slot iRBNER A Boolean variable to indicate whether RBNER (open claims) should be simulated.
#' @slot iROPEN A Boolean variable to indicate whether claim reopen should be simulated.
#' @slot iIBNR A Boolean variable to indicate whether IBNR claims should be simulated.
#' @slot iUPR A Boolean variable to indicate whether future claims should be simulated.
#' @slot fIBNER IBNER development factor.
#' @slot severity Severity distribution.
#' @slot frequency Frequency distribution.
#' @slot reportLag Report lag distribution.
#' @slot settlementLag Settlement lag distribution.
#' @slot reopen Claim reopen probability based on the number of years after settlement till valuation date.
#' @slot reopenLag Reopen lag distribution.
#' @slot resettleLag Resettlement lag distribution.
#' @slot roDevFac Reopened claim development factor.
#' @slot ioDevFac A numeric variable to indicate the method of loss development for open claim severity. 1: Conditional distribution based on paid loss; 2: conditional distribution based on incurred loss; 3: year-to-year development factors
#' @slot irDevFac A numeric variable to indicate the method of loss development for claim reopen severity simulation. 1: Conditional distribution based on paid loss; 2: conditional distribution based on incurred loss; 3: year-to-year development factors
#' @slot freqIndex Frequency distribution time index.
#' @slot severityIndex Severity distribution time index.
#' @slot exposureIndex Exposure time index for IBNR or UPR.
#' @slot iCopula Whether copula is used to model severity, report lag and settlement lag.
#' @slot ssrCopula Copula object used for severity, report lag and settlement lag.
#' @slot sdata Indicating whether only closed claims (CLOSED) or closed + open claims (ALL) will be used for severity fitting.
#' @slot p0 An yearly table that controls the probability of invalid claim, excluding these valid claims less than deductible based on development year. It is based on the DevFac class.
setClass("ClaimType",
slots=c(
simno="character",
line="character",
claimType="character",
# claimClass="character",
iRBNER="logical",
iROPEN="logical",
iIBNR="logical",
iUPR="logical",
fIBNER="DevFac",
severity="Distribution",
frequency="Distribution",
reportLag="Distribution",
settlementLag="Distribution",
reopen="DevFac",
reopenLag="Distribution",
resettleLag="Distribution",
roDevFac="DevFac",
ioDevFac="numeric",
irDevFac="numeric",
IBNRfreqIndex="Index",
UPRfreqIndex="Index",
severityIndex="Index",
exposureIndex="Index",
iCopula="logical",
ssrCopula="CopulaObj",
laeDevFac="DevFac",
deductible="Distribution",
limit="Distribution",
sdata="character",
p0="DevFac"
),
prototype=list(
simno="XXX",
line="Line1",
claimType="Normal",
# claimClass="RBNER",
iRBNER=TRUE,
iROPEN=TRUE,
iIBNR=TRUE,
iUPR=TRUE,
fIBNER=new("DevFac",meanList=c(1.2,1.15,1.1,1.05,1),volList=c(runif(4)/10,0)),
severity=new("Lognormal",p1=3,p2=2),
frequency=new("Poisson",p1=100),
reportLag=new("Exponential",p1=0.25),
settlementLag=new("Exponential",p1=0.1),
reopen=new("DevFac",meanList=c(0.01,0.005,0.002,0.001,0),volList=rep(0,5)),
reopenLag=new("Exponential",p1=0.005),
resettleLag=new("Exponential",p1=0.05),
roDevFac=new("DevFac",meanList=c(1.05,1.1,1,1,1),volList=c(runif(4)/100,0)),
ioDevFac=3,
irDevFac=3,
IBNRfreqIndex=new("Index",monthlyIndex=rep(0,360)),
UPRfreqIndex=new("Index",monthlyIndex=rep(0,360)),
severityIndex=new("Index",monthlyIndex=rep(1,360)),
exposureIndex=new("Index",monthlyIndex=rep(1,360)),
iCopula=FALSE,
ssrCopula=new("CopulaObj",param=c(0,0,0),marginal=list(dist1=new("Lognormal",p1=3,p2=2),dist2=new("Exponential",p1=0.25),dist3=new("Exponential",p1=0.1)),dimension=3),
laeDevFac=new("DevFac",FacModel=TRUE,paras=c(0.1,0.01,0.005,0.002,5)),
deductible=new("Empirical",empirical=matrix(c(0,1,0,0),2,2)),
limit=new("Empirical",empirical=matrix(c(0,1,1e10,1e10),2,2)),
sdata="CLOSED",
p0=new("DevFac",meanList=c(0,0),volList=c(0,0))
)
)
#' Claim simulation at line/type/status level
#' @name claimSample
#' @param object ClaimType object
#' @param ... Additional parameters that may or may not be used.
#' @examples
#' #run time is about 12s(>10s) and is commented out here to avoid long waiting time
#' #library(cascsim)
#' #data(claimdata)
#' ##IBNR simulation
#' #claimobj <- new("ClaimType", line="Auto",claimType="N",iRBNER=FALSE,iROPEN=FALSE,
#' #iIBNR=TRUE,iUPR=FALSE,
#' #IBNRfreqIndex=new("Index",startDate=as.Date("2016-01-01"),
#' #monthlyIndex=rep(30,12)),iCopula=TRUE)
#' #ibnrdata <- claimSample(claimobj,claimdata)
#' #ibnrdata
#' @rdname claimSample-methods
#' @exportMethod claimSample
setGeneric("claimSample", function(object, ...) standardGeneric("claimSample"))
#' @param claimData claim data including existing claims for RBNER and claim reopenness analysis;
#' @param startDate Date from which claim data is included in the analysis;
#' @param evaluationDate Date of evaluation.
#' @rdname claimSample-methods
#' @aliases claimSample,ANY-method
setMethod("claimSample", signature("ClaimType"), function(object, claimData=data.frame(), startDate=as.Date("2012-01-01"), evaluationDate=as.Date("2016-12-31"))
{
tryCatch({
startDate <- toDate(startDate)
evaluationDate <- toDate(evaluationDate)
if (nrow(claimData)>0) {
claimData <- claimData[claimData[,"Type"] == object@claimType & claimData[,"LoB"] == object@line,]
claimData[,"status"] = toupper(claimData[,"status"])
claimData[,"claimLiability"] = toupper(claimData[,"claimLiability"])
claimData[,"occurrenceDate"] = as.character(toDate(claimData[,"occurrenceDate"]))
claimData[,"reportDate"] = as.character(toDate(claimData[,"reportDate"]))
claimData[,"settlementDate"] = as.character(toDate(claimData[,"settlementDate"]))
claimData <- claimData[as.Date(as.character(claimData[,"occurrenceDate"]))>=as.Date(as.character(startDate)),]
claimData <- claimData[as.Date(claimData[,"reportDate"])<=as.Date(evaluationDate),]
}
nid <- 0
sobj <- object@severity
#sobj@truncated <- FALSE
tmpdata = data.frame(ClaimID=character(),
LoB=character(),
Type=character(),
status=character(),
occurrenceDate=character(),
reportDate=character(),
incurredLoss=double(),
osRatio=double(),
settlementDate=character(),
totalLoss=double(),
ultimateLoss=double(),
Deductible=double(),
Limit=double(),
Paid=double(),
LAE=double(),
claimLiability=character(),
expectedLoss=double(),
ultimateLAE=double(),
expectedLAE=double(),
reopenDate=character(),
resettleDate=character(),
reopenLoss=double(),
Sim=double(),
stringsAsFactors=FALSE)
if (object@iRBNER == TRUE) {
if(nrow(claimData)==0){
nobs<-0
} else {
tmpdata <- claimData[claimData[,"status"] == "OPEN" | as.Date(claimData[,"settlementDate"])>as.Date(evaluationDate),]
nobs <- nrow(tmpdata)
}
if (nobs > 0){
occurrenceDates<-as.character(tmpdata[,"occurrenceDate"])
reportDates<-as.character(tmpdata[,"reportDate"])
osRatios<-as.numeric(tmpdata[,"osRatio"])
incurredLosses<-as.numeric(tmpdata[,"incurredLoss"])
settlementDates<-as.character(as.Date(reportDates) + as.numeric(lapply(setMin(object@settlementLag, as.numeric(as.Date(evaluationDate)-as.Date(reportDates))),doSample, n=1)) + as.numeric(as.Date(evaluationDate+1)-as.Date(reportDates))+1)
sindex<-getIndex(object@severityIndex,settlementDates)/getIndex(object@severityIndex,evaluationDate)
tmpdata[,"settlementDate"]<-settlementDates
developmentYears<-ceiling(as.numeric(as.Date(evaluationDate) - as.Date(tmpdata[,"occurrenceDate"]))/365)
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(settlementDates) - as.Date(tmpdata[,"occurrenceDate"]))/365))
if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "exponential" & object@ioDevFac==3){
eibnerdev<-object@fIBNER@paras[1]+object@fIBNER@paras[2]*developmentYears+object@fIBNER@paras[3]*incurredLosses+object@fIBNER@paras[4]*osRatios
ibnerdev<-eibnerdev+rnorm(nobs,0,object@fIBNER@paras[length(object@fIBNER@paras)])
if (length(object@fIBNER@xname)>0){
for (i in c(1:length(object@fIBNER@xname))){
ibnerdev<-ibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
eibnerdev<-eibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
}
}
ultimateLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(log(ibnerdev),0)
expectedLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(log(eibnerdev),0)
} else if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "log" & object@ioDevFac==3){
eibnerdev<-object@fIBNER@paras[1]+object@fIBNER@paras[2]*developmentYears+object@fIBNER@paras[3]*incurredLosses+object@fIBNER@paras[4]*osRatios
ibnerdev<-eibnerdev+rnorm(nobs,0,object@fIBNER@paras[length(object@fIBNER@paras)])
if (length(object@fIBNER@xname)>0){
for (i in c(1:length(object@fIBNER@xname))){
ibnerdev<-ibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
eibnerdev<-eibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
}
}
ultimateLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(exp(ibnerdev),0)
expectedLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(exp(eibnerdev),0)
} else if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "inverse" & object@ioDevFac==3){
eibnerdev<-object@fIBNER@paras[1]+object@fIBNER@paras[2]*developmentYears+object@fIBNER@paras[3]*incurredLosses+object@fIBNER@paras[4]*osRatios
ibnerdev<-eibnerdev+rnorm(nobs,0,object@fIBNER@paras[length(object@fIBNER@paras)])
if (length(object@fIBNER@xname)>0){
for (i in c(1:length(object@fIBNER@xname))){
ibnerdev<-ibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
eibnerdev<-eibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
}
}
ultimateLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(1/ibnerdev,0)
expectedLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(1/eibnerdev,0)
} else if (object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "identity" & object@ioDevFac==3) {
eibnerdev<-object@fIBNER@paras[1]+object@fIBNER@paras[2]*developmentYears+object@fIBNER@paras[3]*incurredLosses+object@fIBNER@paras[4]*osRatios
ibnerdev<-eibnerdev+rnorm(nobs,0,object@fIBNER@paras[length(object@fIBNER@paras)])
if (length(object@fIBNER@xname)>0){
for (i in c(1:length(object@fIBNER@xname))){
ibnerdev<-ibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
eibnerdev<-eibnerdev+object@fIBNER@paras[4+i]*as.numeric(tmpdata[,object@fIBNER@xname[i]])
}
}
ultimateLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(ibnerdev,0)
expectedLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(eibnerdev,0)
} else if (object@ioDevFac==3){
years <- cbind(developmentYears, settlementYears)
ibnerdev<-ultiDevFac(years,meanDevFac=object@fIBNER@meanList,sdDevFac=object@fIBNER@volList,distType=object@fIBNER@distType)
eibnerdev<-ultiDevFac(years,meanDevFac=object@fIBNER@meanList,distType=object@fIBNER@distType)
ultimateLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(ibnerdev,0)
expectedLosses<-as.numeric(tmpdata[,"incurredLoss"])*pmax(eibnerdev,0)
} else if (object@ioDevFac==2){
tmpdata[,"Limit"] <- ifelse(is.na(tmpdata[,"Limit"]),1e10,tmpdata[,"Limit"])
tmpdata[,"Deductible"] <- ifelse(is.na(tmpdata[,"Deductible"]),0,tmpdata[,"Deductible"])
ultimateLosses<-pmax(as.numeric(tmpdata[,"incurredLoss"]),pmax((pmin(tmpdata[,"Limit"]+tmpdata[,"Deductible"],doSample(sobj,nobs))-tmpdata[,"Deductible"]),0))
basicmean<-sampleMean(object@severity)
expectedLosses<-rep(basicmean,nobs)
} else {
tmpdata[,"Limit"] <- ifelse(is.na(tmpdata[,"Limit"]),1e10,tmpdata[,"Limit"])
tmpdata[,"Deductible"] <- ifelse(is.na(tmpdata[,"Deductible"]),0,tmpdata[,"Deductible"])
if(object@iCopula==FALSE){
probs <- Probability(sobj,tmpdata[,"Paid"])
newprobs <- probs + (1-probs)*runif(nobs)
ultimateLosses<-pmax(as.numeric(tmpdata[,"Paid"]),pmax((pmin(tmpdata[,"Limit"]+tmpdata[,"Deductible"],Quantile(sobj,newprobs))-tmpdata[,"Deductible"]),0))
} else {
for (j in c(1:nrow(tmpdata))) {
settlemin <- as.numeric(as.Date(evaluationDate)-as.Date(tmpdata[j,"reportDate"]))
sobjtrunc <- object@ssrCopula@marginal[[1]]@truncated
sobjmin <- object@ssrCopula@marginal[[1]]@min
robjtrunc <- object@ssrCopula@marginal[[2]]@truncated
robjmin <- object@ssrCopula@marginal[[2]]@min
object@ssrCopula@marginal[[1]]@min <- tmpdata[j,"Paid"]
object@ssrCopula@marginal[[1]]@truncated <- TRUE
object@ssrCopula@marginal[[2]]@min <- settlemin
object@ssrCopula@marginal[[2]]@truncated <- TRUE
samples <- copulaSample(object@ssrCopula,1)
tmpdata[j,"settlementDate"] <- as.character(as.Date(as.Date(tmpdata[j,"reportDate"]) + round(samples[,2])))
tmpdata[j,"ultimateLoss"] <- samples[,1]
}
object@ssrCopula@marginal[[1]]@min <- sobjmin
object@ssrCopula@marginal[[1]]@truncated <- sobjtrunc
object@ssrCopula@marginal[[2]]@min <- robjmin
object@ssrCopula@marginal[[2]]@truncated <- robjtrunc
settlementDates <- tmpdata[,"settlementDate"]
ultimateLosses <- tmpdata[,"ultimateLoss"]
}
basicmean<-sampleMean(object@severity)
expectedLosses<-rep(basicmean,nobs)
}
tmpdata[,"Limit"] <- ifelse(is.na(tmpdata[,"Limit"]),1e10,tmpdata[,"Limit"])
tmpdata[,"ultimateLoss"]<-pmin(round(ultimateLosses*sindex,2),tmpdata[,"Limit"])
zeros<-simP0(settlementYears,object@p0@meanList)
exzeros<-expectZeros(settlementYears,object@p0@meanList)
tmpdata[,"expectedLoss"]<-pmin(round(expectedLosses*sindex,2),tmpdata[,"Limit"])*(1-exzeros)
tmpdata[,"claimLiability"] <- ifelse(zeros==0,FALSE,TRUE)
tmpdata[,"ultimateLoss"]<- ifelse(tmpdata[,"claimLiability"] == FALSE, 0, tmpdata[,"ultimateLoss"])
if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "exponential" & object@ioDevFac==3){
laes<-as.numeric(tmpdata[,"LAE"])*pmax(log(ibnerdev)*sindex,0)
elaes<-as.numeric(tmpdata[,"LAE"])*pmax(log(eibnerdev)*sindex,0)
} else if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "log" & object@ioDevFac==3){
laes<-as.numeric(tmpdata[,"LAE"])*pmax(exp(ibnerdev)*sindex,0)
elaes<-as.numeric(tmpdata[,"LAE"])*pmax(exp(eibnerdev)*sindex,0)
} else if(object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "inverse" & object@ioDevFac==3){
laes<-as.numeric(tmpdata[,"LAE"])*pmax(1/ibnerdev*sindex,0)
elaes<-as.numeric(tmpdata[,"LAE"])*pmax(1/eibnerdev*sindex,0)
} else if (object@fIBNER@FacModel==TRUE & object@fIBNER@fun == "identity" & object@ioDevFac==3) {
laes<-as.numeric(tmpdata[,"LAE"])*pmax(ibnerdev*sindex,0)
elaes<-as.numeric(tmpdata[,"LAE"])*pmax(eibnerdev*sindex,0)
} else if (object@ioDevFac==3){
laes<-as.numeric(tmpdata[,"LAE"])*pmax(ibnerdev*sindex,0)
elaes<-as.numeric(tmpdata[,"LAE"])*pmax(eibnerdev*sindex,0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "exponential" & object@ioDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(log(laedev),0)
elaes<-pmax(log(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "log" & object@ioDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(exp(laedev),0)
elaes<-pmax(exp(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "inverse" & object@ioDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(1/laedev,0)
elaes<-pmax(1/elaedev,0)
} else if (object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "identity" & object@ioDevFac<3) {
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(tmpdata[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(laedev,0)
elaes<-pmax(elaedev,0)
} else {
years <-cbind(developmentYears, settlementYears)
laedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList,sdDevFac=object@laeDevFac@volList, distType=object@laeDevFac@distType)
elaedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList, distType=object@laeDevFac@distType)
laes<-as.numeric(tmpdata[,"ultimateLoss"])*pmax(laedev,0)
elaes<-as.numeric(tmpdata[,"expectedLoss"])*pmax(elaedev,0)
}
tmpdata[,"ultimateLAE"]<-round(laes,2)#*sindex
tmpdata[,"expectedLAE"]<-round(elaes,2)#*sindex
tmpdata[,"reopenDate"]<-NA
tmpdata[,"resettleDate"]<-NA
tmpdata[,"reopenLoss"]<-NA
tmpdata[,"Sim"]<-object@simno
tmpdata<-tmpdata[,c("ClaimID","LoB","Type","status","occurrenceDate","reportDate","incurredLoss","osRatio","settlementDate","totalLoss","ultimateLoss","Deductible",
"Limit","Paid","LAE","claimLiability","expectedLoss","ultimateLAE","expectedLAE","reopenDate","resettleDate","reopenLoss","Sim")]
#return(claimData)
} else {
tmpdata = data.frame(ClaimID=character(),
LoB=character(),
Type=character(),
status=character(),
occurrenceDate=character(),
reportDate=character(),
incurredLoss=double(),
osRatio=double(),
settlementDate=character(),
totalLoss=double(),
ultimateLoss=double(),
Deductible=double(),
Limit=double(),
Paid=double(),
LAE=double(),
claimLiability=character(),
expectedLoss=double(),
ultimateLAE=double(),
expectedLAE=double(),
reopenDate=character(),
resettleDate=character(),
reopenLoss=double(),
Sim=double(),
stringsAsFactors=FALSE)
#return(noData)
}
}
if (object@iROPEN == TRUE) {
if(nrow(claimData)==0){
nobs<-0
} else {
claimData <- claimData[claimData[,"status"] == "CLOSED",]# | as.Date(claimData[,"settlementDate"])<=as.Date(evaluationDate),]
nobs <- nrow(claimData)
}
if(nobs>0){
settlementDates<-as.Date(claimData[,"settlementDate"])
occurrenceDates<-as.character(claimData[,"occurrenceDate"])
reportDates<-as.character(claimData[,"reportDate"])
osRatios<-as.numeric(claimData[,"osRatio"])
incurredLosses<-as.numeric(claimData[,"incurredLoss"])
closelags <- as.numeric(as.Date(evaluationDate) - as.Date(claimData[,"settlementDate"]))
closeyears <- pmax(1,ceiling(closelags/365))
developmentYears<-ceiling(as.numeric(as.Date(evaluationDate) - as.Date(claimData[,"occurrenceDate"]))/365)
if(object@reopen@FacModel==TRUE & object@reopen@fun == "exponential"){
reopenp<-object@reopen@paras[1]+object@reopen@paras[2]*developmentYears+object@reopen@paras[3]*incurredLosses+object@reopen@paras[4]*osRatios+rnorm(nobs,0,object@reopen@paras[length(object@reopen@paras)])
if (length(object@reopen@xname)>0){
for (i in c(1:length(object@reopen@xname))){
reopenp<-reopenp+object@reopen@paras[4+i]*as.numeric(claimData[,object@reopen@paras[i]])
}
}
reopens <- ifelse(runif(length(closeyears))<=log(reopenp),1,0)
reopenDates <- ifelse(reopens == 0, NA, as.character(settlementDates + as.numeric(lapply(setMin(object@reopenLag, as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))),doSample, n=1)) + as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))) + 1)
resettleDates <- ifelse(reopens == 0, NA, as.character(reopenDates + as.numeric(lapply(object@resettleLag,doSample, n=1))))
sindex<-getIndex(object@severityIndex,resettleDates)/getIndex(object@severityIndex,as.character(settlementDates))
} else if(object@reopen@FacModel==TRUE & object@reopen@fun == "log"){
reopenp<-object@reopen@paras[1]+object@reopen@paras[2]*developmentYears+object@reopen@paras[3]*incurredLosses+object@reopen@paras[4]*osRatios+rnorm(nobs,0,object@reopen@paras[length(object@reopen@paras)])
if (length(object@reopen@xname)>0){
for (i in c(1:length(object@reopen@xname))){
reopenp<-reopenp+object@reopen@paras[4+i]*as.numeric(claimData[,object@reopen@paras[i]])
}
}
reopens <- ifelse(runif(length(closeyears))<=exp(reopenp),1,0)
reopenDates <- ifelse(reopens == 0, NA, as.character(settlementDates + as.numeric(lapply(setMin(object@reopenLag, as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))),doSample, n=1)) + as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))) + 1)
resettleDates <- ifelse(reopens == 0, NA, as.character(reopenDates + as.numeric(lapply(object@resettleLag,doSample, n=1))))
sindex<-getIndex(object@severityIndex,resettleDates)/getIndex(object@severityIndex,as.character(settlementDates))
} else if(object@reopen@FacModel==TRUE & object@reopen@fun == "inverse"){
reopenp<-object@reopen@paras[1]+object@reopen@paras[2]*developmentYears+object@reopen@paras[3]*incurredLosses+object@reopen@paras[4]*osRatios+rnorm(nobs,0,object@reopen@paras[length(object@reopen@paras)])
if (length(object@reopen@xname)>0){
for (i in c(1:length(object@reopen@xname))){
reopenp<-reopenp+object@reopen@paras[4+i]*as.numeric(claimData[,object@reopen@paras[i]])
}
}
reopens <- ifelse(runif(length(closeyears))<=1/reopenp,1,0)
reopenDates <- ifelse(reopens == 0, NA, as.character(settlementDates + as.numeric(lapply(setMin(object@reopenLag, as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))),doSample, n=1)) + as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))) + 1)
resettleDates <- ifelse(reopens == 0, NA, as.character(reopenDates + as.numeric(lapply(object@resettleLag,doSample, n=1))))
sindex<-getIndex(object@severityIndex,resettleDates)/getIndex(object@severityIndex,as.character(settlementDates))
} else if (object@reopen@FacModel==TRUE & object@reopen@fun == "identity") {
reopenp<-object@reopen@paras[1]+object@reopen@paras[2]*developmentYears+object@reopen@paras[3]*incurredLosses+object@reopen@paras[4]*osRatios+rnorm(nobs,0,object@reopen@paras[length(object@reopen@paras)])
if (length(object@reopen@xname)>0){
for (i in c(1:length(object@reopen@xname))){
reopenp<-reopenp+object@reopen@paras[4+i]*as.numeric(claimData[,object@reopen@paras[i]])
}
}
reopens <- ifelse(runif(length(closeyears))<=reopenp,1,0)
reopenDates <- ifelse(reopens == 0, NA, as.character(settlementDates + as.numeric(lapply(setMin(object@reopenLag, as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))),doSample, n=1))+as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))+1))
resettleDates <- ifelse(reopens == 0, NA, as.character(reopenDates + as.numeric(lapply(object@resettleLag,doSample, n=1))))
sindex<-getIndex(object@severityIndex,resettleDates)/getIndex(object@severityIndex,as.character(settlementDates))
} else {
reopens<-rreopen(closeyears,object@reopen@meanList)
reopenDates <- ifelse(reopens == 0, NA, as.character(settlementDates + as.numeric(lapply(setMin(object@reopenLag, as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))),doSample, n=1))+as.numeric(as.Date(evaluationDate)-as.Date(settlementDates))+1))
resettleDates <- ifelse(reopens == 0, NA, as.character(as.Date(reopenDates) + as.numeric(doSample(object@resettleLag, n=length(reopens)))))
sindex<-getIndex(object@severityIndex,resettleDates)/getIndex(object@severityIndex,as.character(settlementDates))
}
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(resettleDates) - as.Date(claimData[,"occurrenceDate"]))/365))
if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "exponential" & object@irDevFac==3){
erodev<-object@roDevFac@paras[1]+object@roDevFac@paras[2]*developmentYears+object@roDevFac@paras[3]*incurredLosses+object@roDevFac@paras[4]*osRatios
rodev<-erodev+rnorm(nobs,0,object@roDevFac@paras[length(object@roDevFac@paras)])
if (length(object@roDevFac@xname)>0){
for (i in c(1:length(object@roDevFac@xname))){
rodev<-rodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
erodev<-erodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
}
}
ultimateLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(log(rodev),0)
expectedLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(log(erodev),0)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses*sindex)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses*sindex)
} else if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "log" & object@irDevFac==3){
erodev<-object@roDevFac@paras[1]+object@roDevFac@paras[2]*developmentYears+object@roDevFac@paras[3]*incurredLosses+object@roDevFac@paras[4]*osRatios
rodev<-erodev+rnorm(nobs,0,object@roDevFac@paras[length(object@roDevFac@paras)])
if (length(object@roDevFac@xname)>0){
for (i in c(1:length(object@roDevFac@xname))){
rodev<-rodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
erodev<-erodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
}
}
ultimateLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(exp(rodev),0)
expectedLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(exp(erodev),0)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses*sindex)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses*sindex)
} else if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "inverse" & object@irDevFac==3){
erodev<-object@roDevFac@paras[1]+object@roDevFac@paras[2]*developmentYears+object@roDevFac@paras[3]*incurredLosses+object@roDevFac@paras[4]*osRatios
rodev<-erodev+rnorm(nobs,0,object@roDevFac@paras[length(object@roDevFac@paras)])
if (length(object@roDevFac@xname)>0){
for (i in c(1:length(object@roDevFac@xname))){
rodev<-rodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
erodev<-erodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
}
}
ultimateLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(1/rodev,0)
expectedLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(1/erodev,0)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses*sindex)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses*sindex)
} else if (object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "identity" & object@irDevFac==3) {
erodev<-object@roDevFac@paras[1]+object@roDevFac@paras[2]*developmentYears+object@roDevFac@paras[3]*incurredLosses+object@roDevFac@paras[4]*osRatios
rodev<-erodev+rnorm(nobs,0,object@roDevFac@paras[length(object@roDevFac@paras)])
if (length(object@roDevFac@xname)>0){
for (i in c(1:length(object@roDevFac@xname))){
rodev<-rodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
erodev<-erodev+object@roDevFac@paras[4+i]*as.numeric(claimData[,object@roDevFac@xname[i]])
}
}
ultimateLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(rodev,0)
expectedLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(erodev,0)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses*sindex)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses*sindex)
} else if (object@irDevFac==3){
years <- cbind(developmentYears, settlementYears)
rodev<-ultiDevFac(years,meanDevFac=object@roDevFac@meanList,sdDevFac=object@roDevFac@volList,distType=object@roDevFac@distType)
erodev<-ultiDevFac(years,meanDevFac=object@roDevFac@meanList,distType=object@roDevFac@distType)
ultimateLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(rodev,0)
expectedLosses<-as.numeric(claimData[,"incurredLoss"])*pmax(erodev,0)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses*sindex)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses*sindex)
} else if (object@irDevFac==2){
claimData[,"Limit"] <- ifelse(is.na(claimData[,"Limit"]),1e10,claimData[,"Limit"])
claimData[,"Deductible"] <- ifelse(is.na(claimData[,"Deductible"]),0,claimData[,"Deductible"])
ultimateLosses<-pmax(as.numeric(claimData[,"incurredLoss"]),pmax((pmin(claimData[,"Limit"]+claimData[,"Deductible"],doSample(sobj,nobs)*sindex)-claimData[,"Deductible"]),0))
basicmean<-sampleMean(object@severity)
expectedLosses<-rep(basicmean,nobs)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses)
} else {
claimData[,"Limit"] <- ifelse(is.na(claimData[,"Limit"]),1e10,claimData[,"Limit"])
claimData[,"Deductible"] <- ifelse(is.na(claimData[,"Deductible"]),0,claimData[,"Deductible"])
probs <- Probability(sobj,claimData[,"Paid"])
newprobs <- probs + (1-probs)*runif(nobs)
ultimateLosses<-pmax(as.numeric(claimData[,"Paid"]),pmax((pmin(claimData[,"Limit"]+claimData[,"Deductible"],Quantile(sobj,newprobs)*sindex)-claimData[,"Deductible"]),0))
basicmean<-sampleMean(object@severity)
expectedLosses<-rep(basicmean,nobs)
ultimateLosses<- ifelse(reopens == 0, NA, ultimateLosses)
expectedLosses<- ifelse(reopens == 0, NA, expectedLosses)
}
if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "exponential" & object@irDevFac==3){
laes<-as.numeric(claimData[,"LAE"])*pmax(log(rodev),0)
elaes<-as.numeric(claimData[,"LAE"])*pmax(log(erodev),0)
laes<- ifelse(reopens == 0, NA, laes*sindex)
elaes<- ifelse(reopens == 0, NA, elaes*sindex)
} else if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "log" & object@irDevFac==3){
laes<-as.numeric(claimData[,"LAE"])*pmax(exp(rodev),0)
elaes<-as.numeric(claimData[,"LAE"])*pmax(exp(erodev),0)
laes<- ifelse(reopens == 0, NA, laes*sindex)
elaes<- ifelse(reopens == 0, NA, elaes*sindex)
} else if(object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "inverse" & object@irDevFac==3){
laes<-as.numeric(claimData[,"LAE"])*pmax(1/rodev,0)
elaes<-as.numeric(claimData[,"LAE"])*pmax(1/erodev,0)
laes<- ifelse(reopens == 0, NA, laes*sindex)
elaes<- ifelse(reopens == 0, NA, elaes*sindex)
} else if (object@roDevFac@FacModel==TRUE & object@roDevFac@fun == "identity" & object@irDevFac==3) {
laes<-as.numeric(claimData[,"LAE"])*pmax(rodev,0)
elaes<-as.numeric(claimData[,"LAE"])*pmax(erodev,0)
laes<- ifelse(reopens == 0, NA, laes*sindex)
elaes<- ifelse(reopens == 0, NA, elaes*sindex)
} else if (object@irDevFac==3){
laes<-as.numeric(claimData[,"LAE"])*pmax(rodev,0)
elaes<-as.numeric(claimData[,"LAE"])*pmax(erodev,0)
laes<- ifelse(reopens == 0, NA, laes*sindex)
elaes<- ifelse(reopens == 0, NA, elaes*sindex)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "exponential" & object@irDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(log(laedev),0)
elaes<-pmax(log(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "log" & object@irDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(exp(laedev),0)
elaes<-pmax(exp(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "inverse" & object@irDevFac<3){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(1/laedev,0)
elaes<-pmax(1/elaedev,0)
} else if (object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "identity") {
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nobs,0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
if (length(object@laeDevFac@xname)>0){
for (i in c(1:length(object@laeDevFac@xname))){
laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
}
}
laes<-pmax(laedev,0)
elaes<-pmax(elaedev,0)
} else {
years <- cbind(developmentYears, settlementYears)
laedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList,sdDevFac=object@laeDevFac@volList, distType=object@laeDevFac@distType)
elaedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList, distType=object@laeDevFac@distType)
laes<-ultimateLosses*pmax(laedev,0)
elaes<-expectedLosses*pmax(elaedev,0)
}
claimData[,"ultimateLoss"]<-claimData[,"incurredLoss"]
claimData[,"expectedLoss"]<-round(expectedLosses,2)
claimData[,"ultimateLAE"]<-round(laes,2)#*sindex
claimData[,"expectedLAE"]<-round(elaes,2)#*sindex
claimData[,"reopenDate"]<-reopenDates
claimData[,"resettleDate"]<-resettleDates
claimData[,"Limit"] <- ifelse(is.na(claimData[,"Limit"]),1e10,claimData[,"Limit"])
claimData[,"reopenLoss"]<-pmin(round(ultimateLosses,2),claimData[,"Limit"])
claimData[,"Sim"]<-object@simno
claimData<-claimData[,c("ClaimID","LoB","Type","status","occurrenceDate","reportDate","incurredLoss","osRatio","settlementDate","totalLoss","ultimateLoss","Deductible",
"Limit","Paid","LAE","claimLiability","expectedLoss","ultimateLAE","expectedLAE","reopenDate","resettleDate","reopenLoss","Sim")]
tmpdata <- rbind(tmpdata,claimData)
}
}
if (object@iIBNR==TRUE) {
basicmean<-sampleMean(object@severity)
days <- c(31,28,31,30,31,30,31,31,30,31,30,31)
nid <- length(object@IBNRfreqIndex@monthlyIndex)
startyear <- substr(as.character(object@IBNRfreqIndex@startDate),1,4)
startmonth <- substr(as.character(object@IBNRfreqIndex@startDate),6,7)
ibnrdata = data.frame(ClaimID=character(),
LoB=character(),
Type=character(),
status=character(),
occurrenceDate=character(),
reportDate=character(),
incurredLoss=double(),
osRatio=double(),
settlementDate=character(),
totalLoss=double(),
ultimateLoss=double(),
Deductible=double(),
Limit=double(),
Paid=double(),
LAE=double(),
claimLiability=character(),
expectedLoss=double(),
ultimateLAE=double(),
expectedLAE=double(),
reopenDate=character(),
resettleDate=character(),
reopenLoss=double(),
Sim=double(),
stringsAsFactors=FALSE)
if(object@iCopula==TRUE){
ssindex<-vector()
for(i in c(1:nid)){
if(object@IBNRfreqIndex@monthlyIndex[i] > 0) {
currentmonth <- (i-1) %% 12 + as.numeric(startmonth)
currentyear <- floor((i-1)/12)+as.numeric(startyear)
if(currentmonth>12) {
currentmonth <- currentmonth-12
currentyear <- currentyear+1
}
umax <- days[currentmonth]
if(currentyear <10) {
currentmonth <- paste0("0",currentmonth)
}
occurrenceDate<-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01"))+round(runif(object@IBNRfreqIndex@monthlyIndex[i],min=0,max=(umax-1)))
ibnrs<-occurrenceDate
sobjtrunc <- object@ssrCopula@marginal[[1]]@truncated
robjtrunc <- object@ssrCopula@marginal[[3]]@truncated
robjmin <- object@ssrCopula@marginal[[3]]@min
object@ssrCopula@marginal[[1]]@truncated <- FALSE
object@ssrCopula@marginal[[3]]@min <- as.numeric(as.Date(evaluationDate)-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01")))
object@ssrCopula@marginal[[3]]@truncated <- TRUE
samples <- copulaSample(object@ssrCopula,object@IBNRfreqIndex@monthlyIndex[i])
object@ssrCopula@marginal[[1]]@truncated <- sobjtrunc
object@ssrCopula@marginal[[3]]@truncated <- robjtrunc
object@ssrCopula@marginal[[3]]@min <- robjmin
reportDate<-as.Date(occurrenceDate + round(samples[,3]) + as.numeric(as.Date(evaluationDate)-as.Date(occurrenceDate))+1)
settlementDate<-as.Date(reportDate + round(samples[,2]))
sindex<-getIndex(object@severityIndex,settlementDate)/getIndex(object@severityIndex,evaluationDate)
ssindex<-c(ssindex,sindex)
deductibles<-round(doSample(object@deductible,nrow(samples)),-2)
limits<-round(doSample(object@limit,nrow(samples)),-2)
totalLoss<-samples[,1]*sindex
ultimateLoss<-pmax(0,pmin(limits+deductibles,totalLoss)-deductibles)
expectedLoss<-basicmean*sindex
ibnrs<-cbind(as.character(occurrenceDate),as.character(reportDate),as.character(settlementDate),round(ultimateLoss,2),round(expectedLoss,2))
ibnrs<-as.data.frame(ibnrs)
colnames(ibnrs)<-c("occurrenceDate","reportDate","settlementDate","ultimateLoss","expectedLoss")
ibnrs$totalLoss<-totalLoss
ibnrs$ultimateLoss<-as.numeric(as.character(ibnrs$ultimateLoss))
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(settlementDate) - as.Date(occurrenceDate))/365))
zeros<-simP0(settlementYears,object@p0@meanList)
exzeros<-expectZeros(settlementYears,object@p0@meanList)
ibnrs$claimLiability <- ifelse(zeros==0,FALSE,TRUE)
ibnrs$ultimateLoss <- ifelse(ibnrs$claimLiability == FALSE, 0, ibnrs$ultimateLoss)
ibnrs[,"Deductible"]<-deductibles
ibnrs[,"Limit"]<-limits
ibnrs[,"Paid"]<-ibnrs$ultimateLoss
ibnrs$expectedLoss<-as.numeric(as.character(ibnrs$expectedLoss))*(1-exzeros)
ibnrs[,"LoB"]<-object@line
ibnrs[,"Type"]<-object@claimType
ibnrs[,"status"]<-"IBNR"
ibnrs[,"incurredLoss"]<-ibnrs$ultimateLoss
ibnrs[,"osRatio"]<-0
ibnrs[,"ClaimID"]<-as.character(paste0(i,"X",c(1:object@IBNRfreqIndex@monthlyIndex[i])))
ibnrdata<-rbind(ibnrdata,ibnrs)
}
}
} else {
ssindex<-vector()
for(i in c(1:nid)){
if(object@IBNRfreqIndex@monthlyIndex[i] > 0) {
currentmonth <- (i-1) %% 12 + as.numeric(startmonth)
currentyear <- floor((i-1)/12)+as.numeric(startyear)
if(currentmonth>12) {
currentmonth <- currentmonth-12
currentyear <- currentyear+1
}
umax <- days[currentmonth]
if(currentyear <10) {
currentmonth <- paste0("0",currentmonth)
}
occurrenceDate<-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01"))+round(runif(object@IBNRfreqIndex@monthlyIndex[i],min=0,max=(umax-1)))
ibnrs<-occurrenceDate
object@reportLag@min <- as.numeric(as.Date(evaluationDate)-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01")))
object@reportLag@truncated <- TRUE
reportDate<-as.Date(occurrenceDate + round(doSample(object@reportLag,object@IBNRfreqIndex@monthlyIndex[i])) + as.numeric(as.Date(evaluationDate)-as.Date(occurrenceDate))+1)
settlementDate<-as.Date(reportDate + round(doSample(object@settlementLag,object@IBNRfreqIndex@monthlyIndex[i])))
sindex<-getIndex(object@severityIndex,settlementDate)#/getIndex(object@severityIndex,evaluationDate)
ssindex<-c(ssindex,sindex)
deductibles<-round(doSample(object@deductible,object@IBNRfreqIndex@monthlyIndex[i]),-2)
limits<-round(doSample(object@limit,object@IBNRfreqIndex@monthlyIndex[i]),-2)
totalLoss<-doSample(sobj,object@IBNRfreqIndex@monthlyIndex[i])*sindex
ultimateLoss<-pmax(0,pmin(limits+deductibles,totalLoss)-deductibles)
expectedLoss<-basicmean*sindex
ibnrs<-cbind(as.character(occurrenceDate),as.character(reportDate),as.character(settlementDate),round(ultimateLoss,2),round(expectedLoss,2))
ibnrs<-as.data.frame(ibnrs)
colnames(ibnrs)<-c("occurrenceDate","reportDate","settlementDate","ultimateLoss","expectedLoss")
ibnrs$totalLoss<-totalLoss
ibnrs$ultimateLoss<-as.numeric(as.character(ibnrs$ultimateLoss))
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(settlementDate) - as.Date(occurrenceDate))/365))
zeros<-simP0(settlementYears,object@p0@meanList)
exzeros<-expectZeros(settlementYears,object@p0@meanList)
ibnrs$claimLiability <- ifelse(zeros==0,FALSE,TRUE)
ibnrs$ultimateLoss <- ifelse(ibnrs$claimLiability == FALSE, 0, ibnrs$ultimateLoss)
ibnrs[,"Deductible"]<-deductibles
ibnrs[,"Limit"]<-limits
ibnrs[,"Paid"]<-ibnrs$ultimateLoss
ibnrs$expectedLoss<-as.numeric(as.character(ibnrs$expectedLoss))*(1-exzeros)
ibnrs[,"LoB"]<-object@line
ibnrs[,"Type"]<-object@claimType
ibnrs[,"status"]<-"IBNR"
ibnrs[,"incurredLoss"]<-ibnrs$ultimateLoss
ibnrs[,"osRatio"]<-0
ibnrs[,"ClaimID"]<-as.character(paste0(i,"X",c(1:object@IBNRfreqIndex@monthlyIndex[i])))
ibnrdata<-rbind(ibnrdata,ibnrs)
}
}
}
if (nrow(ibnrdata)>0) {
developmentYears<-ceiling(as.numeric(as.Date(ibnrdata[,"reportDate"]) - as.Date(ibnrdata[,"occurrenceDate"]))/365)
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(ibnrdata[,"settlementDate"]) - as.Date(ibnrdata[,"occurrenceDate"]))/365))
incurredLosses<-ibnrdata[,"ultimateLoss"]
osRatios<-ibnrdata[,"osRatio"]
if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "exponential"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(ibnrdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(log(laedev),0)
elaes<-pmax(log(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "log"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(ibnrdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(exp(laedev),0)
elaes<-pmax(exp(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "inverse"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(ibnrdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(1/laedev,0)
elaes<-pmax(1/elaedev,0)
} else if (object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "identity") {
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(ibnrdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(laedev,0)
elaes<-pmax(elaedev,0)
} else {
years <- cbind(developmentYears,settlementYears)
laedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList,sdDevFac=object@laeDevFac@volList, distType=object@laeDevFac@distType)
elaedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList, distType=object@laeDevFac@distType)
laes<-as.numeric(ibnrdata[,"ultimateLoss"])*pmax(laedev,0)
elaes<-as.numeric(ibnrdata[,"expectedLoss"])*pmax(elaedev,0)
}
ibnrdata[,"ultimateLAE"]<-round(laes,2)#*ssindex
ibnrdata[,"expectedLAE"]<-round(elaes,2)#*ssindex
ibnrdata[,"reopenDate"]<-NA
ibnrdata[,"resettleDate"]<-NA
ibnrdata[,"reopenLoss"]<-NA
ibnrdata[,"LAE"]<-ibnrdata[,"ultimateLAE"]*1
ibnrdata[,"Sim"]<-object@simno
# ibnrdata<-ibnrdata[,c(11,6,7,8,1,2,9,10,3,4,17,5,12:16,18)]
ibnrdata<-ibnrdata[,c("ClaimID","LoB","Type","status","occurrenceDate","reportDate","incurredLoss","osRatio","settlementDate","totalLoss","ultimateLoss","Deductible",
"Limit","Paid","LAE","claimLiability","expectedLoss","ultimateLAE","expectedLAE","reopenDate","resettleDate","reopenLoss","Sim")]
tmpdata <- rbind(tmpdata,ibnrdata)
}
}
if (object@iUPR == TRUE) {
basicmean<-sampleMean(object@severity)
days <- c(31,28,31,30,31,30,31,31,30,31,30,31)
nid <- length(object@UPRfreqIndex@monthlyIndex)
startyear <- substr(as.character(object@UPRfreqIndex@startDate),1,4)
startmonth <- substr(as.character(object@UPRfreqIndex@startDate),6,7)
uprdata = data.frame(ClaimID=character(),
LoB=character(),
Type=character(),
status=character(),
occurrenceDate=character(),
reportDate=character(),
incurredLoss=double(),
osRatio=double(),
settlementDate=character(),
totalLoss=double(),
ultimateLoss=double(),
Deductible=double(),
Limit=double(),
Paid=double(),
LAE=double(),
claimLiability=character(),
expectedLoss=double(),
ultimateLAE=double(),
expectedLAE=double(),
reopenDate=character(),
resettleDate=character(),
reopenLoss=double(),
Sim=double(),
stringsAsFactors=FALSE)
if(object@iCopula==TRUE){
ssindex<-vector()
for(i in c(1:nid)){
if(object@UPRfreqIndex@monthlyIndex[i] > 0) {
currentmonth <- (i-1) %% 12 + as.numeric(startmonth)
currentyear <- floor((i-1)/12)+as.numeric(startyear)
if(currentmonth>12) {
currentmonth <- currentmonth-12
currentyear <- currentyear+1
}
umax <- days[currentmonth]
if(currentyear <10) {
currentmonth <- paste0("0",currentmonth)
}
occurrenceDate<-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01"))+round(runif(object@UPRfreqIndex@monthlyIndex[i],min=0,max=(umax-1)))
uprs<-occurrenceDate
sobjtrunc <- object@ssrCopula@marginal[[1]]@truncated
object@ssrCopula@marginal[[1]]@truncated <- FALSE
samples <- copulaSample(object@ssrCopula,object@UPRfreqIndex@monthlyIndex[i])
object@ssrCopula@marginal[[1]]@truncated <- sobjtrunc
reportDate<-as.Date(occurrenceDate + round(samples[,3]))# + as.numeric(as.Date(evaluationDate)-as.Date(occurrenceDate))+1)
settlementDate<-as.Date(reportDate + round(samples[,2]))
sindex<-getIndex(object@severityIndex,settlementDate)#/getIndex(object@severityIndex,evaluationDate)
ssindex<-c(ssindex,sindex)
deductibles<-round(doSample(object@deductible,nrow(samples)),-2)
limits<-round(doSample(object@limit,nrow(samples)),-2)
totalLoss<-samples[,1]*sindex
ultimateLoss<-pmax(0,pmin(limits+deductibles,totalLoss)-deductibles)
expectedLoss<-basicmean*sindex
uprs<-cbind(as.character(occurrenceDate),as.character(reportDate),as.character(settlementDate),round(ultimateLoss,2),round(expectedLoss,2))
uprs<-as.data.frame(uprs)
colnames(uprs)<-c("occurrenceDate","reportDate","settlementDate","ultimateLoss","expectedLoss")
uprs$totalLoss<-totalLoss
uprs$ultimateLoss<-as.numeric(as.character(uprs$ultimateLoss))
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(settlementDate) - as.Date(occurrenceDate))/365))
zeros<-simP0(settlementYears,object@p0@meanList)
exzeros<-expectZeros(settlementYears,object@p0@meanList)
uprs$claimLiability <- ifelse(zeros==0,FALSE,TRUE)
uprs$ultimateLoss <- ifelse(uprs$claimLiability == FALSE, 0, uprs$ultimateLoss)
uprs[,"Deductible"]<-deductibles
uprs[,"Limit"]<-limits
uprs[,"Paid"]<- uprs$ultimateLoss
uprs$expectedLoss<-as.numeric(as.character(uprs$expectedLoss))*(1-exzeros)
uprs[,"LoB"]<-object@line
uprs[,"Type"]<-object@claimType
uprs[,"status"]<-"UPR"
uprs[,"incurredLoss"]<-uprs$ultimateLoss
uprs[,"osRatio"]<-0
uprs[,"ClaimID"]<-as.character(paste0(i,"X",c(1:object@UPRfreqIndex@monthlyIndex[i])))
uprdata<-rbind(uprdata,uprs)
}
}
} else {
ssindex<-vector()
for(i in c(1:nid)){
if(object@UPRfreqIndex@monthlyIndex[i] > 0) {
currentmonth <- (i-1) %% 12 + as.numeric(startmonth)
currentyear <- floor((i-1)/12)+as.numeric(startyear)
if(currentmonth>12) {
currentmonth <- currentmonth-12
currentyear <- currentyear+1
}
umax <- days[currentmonth]
if(currentyear <10) {
currentmonth <- paste0("0",currentmonth)
}
occurrenceDate<-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01"))+round(runif(object@UPRfreqIndex@monthlyIndex[i],min=0,max=(umax-1)))
uprs<-occurrenceDate
# object@reportLag@min <- as.numeric(as.Date(evaluationDate)-as.Date(paste0(as.character(currentyear),"-",as.character(currentmonth),"-01")))
# object@reportLag@truncated <- TRUE
reportDate<-as.Date(occurrenceDate + round(doSample(object@reportLag,object@UPRfreqIndex@monthlyIndex[i])))
settlementDate<-as.Date(reportDate + round(doSample(object@settlementLag,object@UPRfreqIndex@monthlyIndex[i])))
sindex<-getIndex(object@severityIndex,settlementDate)/getIndex(object@severityIndex,evaluationDate)
ssindex<-c(ssindex,sindex)
deductibles<-round(doSample(object@deductible,object@UPRfreqIndex@monthlyIndex[i]),-2)
limits<-round(doSample(object@limit,object@UPRfreqIndex@monthlyIndex[i]),-2)
totalLoss<-doSample(sobj,object@UPRfreqIndex@monthlyIndex[i])*sindex
ultimateLoss<-pmax(0,pmin(limits+deductibles,totalLoss)-deductibles)
expectedLoss<-basicmean*sindex
uprs<-cbind(as.character(occurrenceDate),as.character(reportDate),as.character(settlementDate),round(ultimateLoss,2),round(expectedLoss,2))
uprs<-as.data.frame(uprs)
colnames(uprs)<-c("occurrenceDate","reportDate","settlementDate","ultimateLoss","expectedLoss")
uprs$totalLoss<-totalLoss
uprs$ultimateLoss<-as.numeric(as.character(uprs$ultimateLoss))
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(settlementDate) - as.Date(occurrenceDate))/365))
zeros<-simP0(settlementYears,object@p0@meanList)
exzeros<-expectZeros(settlementYears,object@p0@meanList)
uprs$claimLiability <- ifelse(zeros==0,FALSE,TRUE)
uprs$ultimateLoss <- ifelse(uprs$claimLiability == FALSE, 0, uprs$ultimateLoss)
uprs[,"Deductible"]<-deductibles
uprs[,"Limit"]<-limits
uprs[,"Paid"]<-uprs$ultimateLoss
uprs$expectedLoss<-as.numeric(as.character(uprs$expectedLoss))*(1-exzeros)
uprs[,"LoB"]<-object@line
uprs[,"Type"]<-object@claimType
uprs[,"status"]<-"UPR"
uprs[,"incurredLoss"]<-uprs$ultimateLoss
uprs[,"osRatio"]<-0
uprs[,"ClaimID"]<-as.character(paste0(i,"X",c(1:object@UPRfreqIndex@monthlyIndex[i])))
uprdata<-rbind(uprdata,uprs)
}
}
}
if (nrow(uprdata)>0) {
developmentYears<-ceiling(as.numeric(as.Date(uprdata[,"reportDate"]) - as.Date(uprdata[,"occurrenceDate"]))/365)
settlementYears<-pmax(1,ceiling(as.numeric(as.Date(uprdata[,"settlementDate"]) - as.Date(uprdata[,"occurrenceDate"]))/365))
incurredLosses<-uprdata[,"ultimateLoss"]
osRatios<-uprdata[,"osRatio"]
if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "exponential"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(uprdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(log(laedev),0)
elaes<-pmax(log(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "log"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(uprdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(exp(laedev),0)
elaes<-pmax(exp(elaedev),0)
} else if(object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "inverse"){
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(uprdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(1/laedev,0)
elaes<-pmax(1/elaedev,0)
} else if (object@laeDevFac@FacModel==TRUE & object@laeDevFac@fun == "identity") {
elaedev<-object@laeDevFac@paras[1]+object@laeDevFac@paras[2]*developmentYears+object@laeDevFac@paras[3]*incurredLosses+object@laeDevFac@paras[4]*osRatios
laedev<-elaedev+rnorm(nrow(uprdata),0,object@laeDevFac@paras[length(object@laeDevFac@paras)])
# if (length(object@laeDevFac@xname)>0){
# for (i in c(1:length(object@laeDevFac@xname))){
# laedev<-laedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# elaedev<-elaedev+object@laeDevFac@paras[4+i]*as.numeric(claimData[,object@laeDevFac@xname[i]])
# }
# }
laes<-pmax(laedev,0)
elaes<-pmax(elaedev,0)
} else {
years <- cbind(developmentYears,settlementYears)
laedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList,sdDevFac=object@laeDevFac@volList, distType=object@laeDevFac@distType)
elaedev<-ultiDevFac(years,meanDevFac=object@laeDevFac@meanList, distType=object@laeDevFac@distType)
laes<-as.numeric(uprdata[,"ultimateLoss"])*pmax(laedev,0)
elaes<-as.numeric(uprdata[,"expectedLoss"])*pmax(elaedev,0)
}
uprdata[,"ultimateLAE"]<-round(laes,2)#*ssindex
uprdata[,"expectedLAE"]<-round(elaes,2)#*ssindex
uprdata[,"reopenDate"]<-NA
uprdata[,"resettleDate"]<-NA
uprdata[,"reopenLoss"]<-NA
uprdata[,"LAE"]<-uprdata[,"ultimateLAE"]*1
uprdata[,"Sim"]<-object@simno
uprdata<-uprdata[,c("ClaimID","LoB","Type","status","occurrenceDate","reportDate","incurredLoss","osRatio","settlementDate","totalLoss","ultimateLoss","Deductible",
"Limit","Paid","LAE","claimLiability","expectedLoss","ultimateLAE","expectedLAE","reopenDate","resettleDate","reopenLoss","Sim")]
tmpdata <- rbind(tmpdata,uprdata)
}
}
return(tmpdata)
}, error = function(err){
message(paste0(">>>Critical Error for claim simulation: Line-", object@line, " Type-", object@claimType, " Sim No.-", object@simno, ", ",err))
gc()
})
})
Try the cascsim package in your browser
Any scripts or data that you put into this service are public.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.