R/acd-forecastLong.R
In TrungLeVn/SgtAcd: Autoregressive Density Model with Skewed Generalized Student-t distribution
Defines functions
.acdforecast2L
.acdforecastL
.acdforecastSpecL
.acdforecastFitL
acdforecastLong
Documented in
acdforecastLong
#--------------------
# ACDforecast method
#' @export acdforecastLong
#' @title Autoregressive Conditional Density model forecast
#' @description Produce forecast for ACD models, basing on the simulation
#' @param fitORspec Either ACDfit or ACDspec object is allowed
#--------------------
# forecast
acdforecastLong = function(fitORspec, data = NULL, horizon = 10, n.roll = 0, out.sample = 0,
m.sim = 1000, cluster = NULL,
skew0 = NULL, shape10 = NULL,skew20 = NULL, ...)
{
UseMethod("acdforecastLong")
}
.acdforecastFitL = function(fitORspec, horizon = 10, n.roll = 0,
m.sim = 1000, cluster = NULL, ...)
{
fit = fitORspec
ans = .acdforecastL(fit = fit, horizon = horizon,
n.roll = n.roll,
m.sim = m.sim,
cluster = cluster, ...)
return(ans)
}
# switch if we supply model specification to the acdforecast function
.acdforecastSpecL = function(fitORspec, data = NULL, horizon = 10, n.roll = 0, out.sample = 0,
m.sim = 1000,
cluster = NULL, skew0 = NULL, shape0 = NULL, ...)
{
spec = fitORspec
ans = .acdforecast2L(spec = spec, data = data, horizon = horizon,
n.roll = n.roll, out.sample = out.sample,
m.sim = m.sim,
cluster = cluster, skew0 = skew0, shape10 = shape10,shape20 = shape20, ...)
return(ans)
}
setMethod("acdforecastLong", signature(fitORspec = "ACDfit"), .acdforecastFitL)
setMethod("acdforecastLong", signature(fitORspec = "ACDspec"), .acdforecastSpecL)
#------------------
#---------------------------
.acdforecastL = function(fit, horizon = 10, n.roll = 0,
m.sim = 1000, cluster = NULL, rseed = NA, ...)
{
if(is.na(rseed[1])){
sseed = 2706:(2706+m.sim-1)
} else{
if(length(rseed) != m.sim) stop("\uacdforecast-->error: rseed must be of length m.sim!\n")
sseed = rseed[1:m.sim]
}
data = fit@model$modeldata$data
Nor = length(fit@model$modeldata$data)
index = fit@model$modeldata$index
period = fit@model$modeldata$period
n.ahead = 1 #Only allow for one-steap ahead as the long-horizon conditional density is not additive
ns = fit@model$n.start ##ns is the number of periods before the end of dataset that is used to fit the model
if(n.roll>0 && ns ==0) stop("\uacdforecast-->error: Rolling forecast is only available when we have out-of-sample data!\n")
N = Nor - ns #N is the length of data that is used to fit model or data[N] is the last point of fitted data
model = fit@model
ipars = fit@fit$ipars
modelinc = model$modelinc
idx = model$pidx
if( n.roll > ns ) stop("\nacdforecast-->error: n.roll must not be greater than out.sample!")
pars = fit@fit$coef
ipars = fit@fit$ipars
# filter data (check external regressor data - must equal length of origData)
# Generate the 1 extra ahead forecast using filter process with the last residuals and z as based values
if(n.roll > 0){ # if we have out-of-sample in the fit object and we wish to rolling forecast
fcreq = ifelse(ns >= (n.ahead+n.roll+1), n.ahead+n.roll + 1, ns)
tmp = xts::xts(data[1:(N + fcreq)], index[1:(N + fcreq)])
fspec = acdspec(variance.model = list(model = model$vmodel$model,
garchOrder = model$vmodel$garchOrder,
variance.targeting = FALSE),
mean.model = list(armaOrder = model$mmodel$armaOrder,
include.mean = model$mmodel$include.mean,
archm = model$mmodel$archm, skm = model$mmodel$skm, pskm = model$mmodel$pskm, adjm = model$mmodel$adjm),
distribution.model = list(model = model$dmodel$model,
skewOrder = model$dmodel$skewOrder, skewshock = model$dmodel$skewshock,
skewmodel = model$dmodel$skewmodel, volsk = model$dmodel$volsk,
shape1Order = model$dmodel$shape1Order, shape1shock = model$dmodel$shape1shock,
shape1model = model$dmodel$shape1model, volsh1 = model$dmodel$volsh1,
shape2Order = model$dmodel$shape2Order, shape2shock = model$dmodel$shape2shock,
shape2model = model$dmodel$shape2model, volsh2 = model$dmodel$volsh2,
exp.rate=model$sbounds[7]))
fspec = setfixedacd(fspec, as.list(coefacd(fit)))
fspec = setboundsacd(fspec,list(shape1 = fit@model$sbounds[3:4],shape2 = fit@model$sbounds[5:6], skew = fit@model$sbounds[1:2]))
flt = acdfilter(spec = fspec, data = tmp, n.old = N, skew0 = fit@fit$tskew[1], shape10 = fit@fit$tshape[1],shape20 = fit@fit$tshape2[1])
sigmafilter = flt@filter$sigma
resfilter = flt@filter$residuals
zfilter = flt@filter$z
tskewfilter = flt@filter$tskew
tshape1filter = flt@filter$tshape1
tshape2filter =flt@filter$tshape2
tempskewfilter = flt@filter$tempskew
tempshape1filter = flt@filter$tempshape1
tempshape2filter = flt@filter$tempshape2
seriesfor = sigmafor = matrix(NA, nrow = n.roll+1, ncol = n.ahead)
skewnessFor = kurtosisFor = matrix(NA,nrow = n.roll + 1, ncol = n.ahead)
# n.roll x n.ahead (n.ahead=1 generted by model)
# n.ahead>1 by simulation
rownames(seriesfor) = rownames(sigmafor) = rownames(skewnessFor) = rownames(kurtosisFor)= as.character(index[N:(N+n.roll)])
colnames(seriesfor) = colnames(sigmafor) = colnames(skewnessFor) = colnames(kurtosisFor) = paste("T+", 1:n.ahead, sep="")
mx = model$maxOrder #At the moment, the package only allows for maximum lag of 1 in both mean, sigma and higher moment equations.
# Hence mx is always 1
#constmean = unname(coefacd(fit)["mu"])
for(i in 1:(n.roll+1)){
spec = fspec
presig = tail(sigmafilter[1:(N+i-1)], mx)
preskew = tail(tskewfilter[1:(N+i-1)], mx)
preshape1 = tail(tshape1filter[1:(N+i-1)], mx)
preshape2 = tail(tshape2filter[1:(N+i-1)], mx)
prereturns = tail(data[1:(N+i-1)], mx)
sim = acdpath(spec, n.sim = horizon, n.start = 0, m.sim = m.sim,
presigma = presig, preskew = preskew,
preshape1 = preshape1,preshape2 = preshape2, prereturns = prereturns,
preresiduals = NA, rseed = rseed,
cluster = cluster)
seriesfor[i, ] = mean(colSums(sim@path$seriesSim))
sigmafor[i, ] = sqrt(mean(colSums(sim@path$sigmaSim^2)))
skewnessFor[i, ] = PerformanceAnalytics::skewness(colSums(sim@path$seriesSim))
kurtosisFor[i, ] = PerformanceAnalytics::kurtosis(colSums(sim@path$seriesSim),method = "excess")
}
} else{
seriesfor = sigmafor = matrix(NA, nrow = 1, ncol = n.ahead)
skewnessFor = kurtosisFor = matrix(NA,nrow = 1, ncol = n.ahead)
# n.roll x n.ahead (n.ahead=1 generted by model)
# n.ahead>1 by simulation
rownames(seriesfor) = rownames(sigmafor) = rownames(skewnessFor) = rownames(kurtosisFor)= as.character(index[N])
colnames(seriesfor) = colnames(sigmafor) = colnames(skewnessFor) = colnames(kurtosisFor) = paste("T+", 1:n.ahead, sep="")
mx = model$maxOrder
presig = as.numeric(tail(sigmaAcd(fit),mx))
preskew = as.numeric(tail(skew(fit),mx))
preshape1 = as.numeric(tail(shape1(fit),mx))
preshape2 = as.numeric(tail(shape2(fit),mx))
preresiduals = as.numeric(tail(residualsAcd(fit),mx))
prereturns = as.numeric(tail(fit@model$modeldata$data[N],mx))
sim = acdsim(fit, n.sim = horizon, m.sim = m.sim,
presigma = presig, preskew = preskew,
preshape1 = preshape1,preshape2 = preshape2, prereturns = prereturns,
preresiduals = preresiduals, rseed = rseed,
cluster = cluster)
seriesfor[,1] = mean(colSums(sim@simulation$seriesSim))
sigmafor[,1] = sqrt(mean(colSums(sim@simulation$sigmaSim^2)))
skewnessFor[,1] = PerformanceAnalytics::skewness(colSums(sim@simulation$seriesSim))
}
fcst = list()
fcst$n.ahead = n.ahead
fcst$n.roll = n.roll
fcst$N = N+ns
fcst$n.start = ns
fcst$seriesFor = seriesfor
fcst$sigmaFor = sigmafor
fcst$skewnessFor = skewnessFor
fcst$kurtosisFor = kurtosisFor
if(n.roll>0){
model$modeldata$sigma = flt@filter$sigma
model$modeldata$residuals = flt@filter$residuals
model$modeldata$tskew = flt@filter$tskew
model$modeldata$tshape1 = flt@filter$tshape1
model$modeldata$tshape2 = flt@filter$tshape2
model$modeldata$Pskewness = flt@filter$Pskewness
} else{
model = fit@model
}
ans = new("ACDforecast",
forecast = fcst,
model = model)
return(ans)
}
# CONTINUE HERE:
.acdforecast2L = function(spec, data = NULL, n.ahead = 10, n.roll = 0, out.sample = 0,
m.sim = 1000, cluster = NULL, skew0 = NULL,
shape0 = NULL, ...)
{
vmodel = spec@model$vmodel$model
xdata = rugarch:::.extractdata(data)
data = xdata$data
index = xdata$index
period = xdata$period
Nor = length(as.numeric(data))
ns = out.sample
N = Nor - ns
model = spec@model
modelinc = model$modelinc
idx = model$pidx
ipars = model$pars
pars = unlist(model$fixed.pars)
parnames = names(pars)
modelnames = rugarch:::.checkallfixed(spec)
if(is.na(all(match(modelnames, parnames), 1:length(modelnames)))) {
cat("\nacdfilter-->error: parameters names do not match specification\n")
cat("Expected Parameters are: ")
cat(paste(modelnames))
cat("\n")
stop("Exiting", call. = FALSE)
}
fspec = spec
setfixed(fspec)<-as.list(pars)
if( n.roll > ns ) stop("\nacdforecast-->error: n.roll must not be greater than out.sample!")
# filter data (check external regressor data - must equal length of origData)
fcreq = ifelse(ns >= (n.ahead+n.roll), n.ahead+n.roll, ns)
# Generate the 1 extra ahead forecast
if((n.ahead+n.roll)<=ns){
tmp = xts(data[1:(N + fcreq)], index[1:(N + fcreq)])
} else{
tmp = xts(c(data[1:(N + fcreq)],0), c(index[1:(N + fcreq)], index[(N + fcreq)]+1))
}
flt = acdfilter(spec = fspec, data = tmp, n.old = N, skew0 = skew0, shape0 = shape0)
sigmafilter = flt@filter$sigma
resfilter = flt@filter$residuals
zfilter = flt@filter$z
tskewfilter = flt@filter$tskew
tshapefilter = flt@filter$tshape
tempskewfilter = flt@filter$tempskew
tempshapefilter = flt@filter$tempshape
seriesfor = sigmafor = matrix(NA, nrow = n.roll+1, ncol = n.ahead)
skewnessFor = kurtosisFor = matrix(NA,nrow = n.roll + 1, ncol = n.ahead)
# n.roll x n.ahead (n.ahead=1 generted by model)
# n.ahead>1 by simulation
rownames(seriesfor) = rownames(sigmafor) = rownames(skewnessFor) = rownames(kurtosisFor)= as.character(index[N:(N+n.roll)])
colnames(seriesfor) = colnames(sigmafor) = colnames(skewnessFor) = colnames(kurtosisFor) = paste("T+", 1:n.ahead, sep="")
mx = model$maxOrder #At the moment, the package only allows for maximum lag of 1 in both mean, sigma and higher moment equations.
# Hence mx is always 1
#constmean = unname(coefacd(fit)["mu"])
for(i in 1:(n.roll+1)){
spec = fspec
presig = tail(sigmafilter[1:(N+i-1)], mx)
preskew = tail(tskewfilter[1:(N+i-1)], mx)
preshape1 = tail(tshape1filter[1:(N+i-1)], mx)
preshape2 = tail(tshape2filter[1:(N+i-1)], mx)
prereturns = tail(data[1:(N+i-1)], mx)
sim = acdpath(spec, n.sim = horizon, n.start = 0, m.sim = m.sim,
presigma = presig, preskew = preskew,
preshape1 = preshape1,preshape2 = preshape2, prereturns = prereturns,
preresiduals = NA, rseed = rseed,
cluster = cluster)
seriesfor[i, ] = mean(colSums(sim@path$seriesSim))
sigmafor[i, ] = sqrt(mean(colSums(sim@path$sigmaSim^2)))
skewnessFor[i, ] = PerformanceAnalytics::skewness(colSums(sim@path$seriesSim))
kurtosisFor[i, ] = PerformanceAnalytics::kurtosis(colSums(sim@path$seriesSim),method = "excess")
}
fcst = list()
fcst$n.ahead = n.ahead
fcst$n.roll = n.roll
fcst$N = N+ns
fcst$n.start = ns
fcst$seriesFor = seriesfor
fcst$sigmaFor = sigmafor
fcst$skewnessFor = skewnessfor
fcst$kurtosisFor = kurtosisFor
model$modeldata$sigma = flt@filter$sigma
model$modeldata$residuals = flt@filter$residuals
model$modeldata$tskew = flt@filter$tskew
model$modeldata$tshape = flt@filter$tshape
ans = new("ACDforecast",
forecast = fcst,
model = model)
return(ans)
}
TrungLeVn/SgtAcd documentation built on July 25, 2021, 4:49 a.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.
Defines functions .acdforecast2L .acdforecastL .acdforecastSpecL .acdforecastFitL acdforecastLong
Documented in acdforecastLong
#--------------------
# ACDforecast method
#' @export acdforecastLong
#' @title Autoregressive Conditional Density model forecast
#' @description Produce forecast for ACD models, basing on the simulation
#' @param fitORspec Either ACDfit or ACDspec object is allowed
#--------------------
# forecast
acdforecastLong = function(fitORspec, data = NULL, horizon = 10, n.roll = 0, out.sample = 0,
m.sim = 1000, cluster = NULL,
skew0 = NULL, shape10 = NULL,skew20 = NULL, ...)
{
UseMethod("acdforecastLong")
}
.acdforecastFitL = function(fitORspec, horizon = 10, n.roll = 0,
m.sim = 1000, cluster = NULL, ...)
{
fit = fitORspec
ans = .acdforecastL(fit = fit, horizon = horizon,
n.roll = n.roll,
m.sim = m.sim,
cluster = cluster, ...)
return(ans)
}
# switch if we supply model specification to the acdforecast function
.acdforecastSpecL = function(fitORspec, data = NULL, horizon = 10, n.roll = 0, out.sample = 0,
m.sim = 1000,
cluster = NULL, skew0 = NULL, shape0 = NULL, ...)
{
spec = fitORspec
ans = .acdforecast2L(spec = spec, data = data, horizon = horizon,
n.roll = n.roll, out.sample = out.sample,
m.sim = m.sim,
cluster = cluster, skew0 = skew0, shape10 = shape10,shape20 = shape20, ...)
return(ans)
}
setMethod("acdforecastLong", signature(fitORspec = "ACDfit"), .acdforecastFitL)
setMethod("acdforecastLong", signature(fitORspec = "ACDspec"), .acdforecastSpecL)
#------------------
#---------------------------
.acdforecastL = function(fit, horizon = 10, n.roll = 0,
m.sim = 1000, cluster = NULL, rseed = NA, ...)
{
if(is.na(rseed[1])){
sseed = 2706:(2706+m.sim-1)
} else{
if(length(rseed) != m.sim) stop("\uacdforecast-->error: rseed must be of length m.sim!\n")
sseed = rseed[1:m.sim]
}
data = fit@model$modeldata$data
Nor = length(fit@model$modeldata$data)
index = fit@model$modeldata$index
period = fit@model$modeldata$period
n.ahead = 1 #Only allow for one-steap ahead as the long-horizon conditional density is not additive
ns = fit@model$n.start ##ns is the number of periods before the end of dataset that is used to fit the model
if(n.roll>0 && ns ==0) stop("\uacdforecast-->error: Rolling forecast is only available when we have out-of-sample data!\n")
N = Nor - ns #N is the length of data that is used to fit model or data[N] is the last point of fitted data
model = fit@model
ipars = fit@fit$ipars
modelinc = model$modelinc
idx = model$pidx
if( n.roll > ns ) stop("\nacdforecast-->error: n.roll must not be greater than out.sample!")
pars = fit@fit$coef
ipars = fit@fit$ipars
# filter data (check external regressor data - must equal length of origData)
# Generate the 1 extra ahead forecast using filter process with the last residuals and z as based values
if(n.roll > 0){ # if we have out-of-sample in the fit object and we wish to rolling forecast
fcreq = ifelse(ns >= (n.ahead+n.roll+1), n.ahead+n.roll + 1, ns)
tmp = xts::xts(data[1:(N + fcreq)], index[1:(N + fcreq)])
fspec = acdspec(variance.model = list(model = model$vmodel$model,
garchOrder = model$vmodel$garchOrder,
variance.targeting = FALSE),
mean.model = list(armaOrder = model$mmodel$armaOrder,
include.mean = model$mmodel$include.mean,
archm = model$mmodel$archm, skm = model$mmodel$skm, pskm = model$mmodel$pskm, adjm = model$mmodel$adjm),
distribution.model = list(model = model$dmodel$model,
skewOrder = model$dmodel$skewOrder, skewshock = model$dmodel$skewshock,
skewmodel = model$dmodel$skewmodel, volsk = model$dmodel$volsk,
shape1Order = model$dmodel$shape1Order, shape1shock = model$dmodel$shape1shock,
shape1model = model$dmodel$shape1model, volsh1 = model$dmodel$volsh1,
shape2Order = model$dmodel$shape2Order, shape2shock = model$dmodel$shape2shock,
shape2model = model$dmodel$shape2model, volsh2 = model$dmodel$volsh2,
exp.rate=model$sbounds[7]))
fspec = setfixedacd(fspec, as.list(coefacd(fit)))
fspec = setboundsacd(fspec,list(shape1 = fit@model$sbounds[3:4],shape2 = fit@model$sbounds[5:6], skew = fit@model$sbounds[1:2]))
flt = acdfilter(spec = fspec, data = tmp, n.old = N, skew0 = fit@fit$tskew[1], shape10 = fit@fit$tshape[1],shape20 = fit@fit$tshape2[1])
sigmafilter = flt@filter$sigma
resfilter = flt@filter$residuals
zfilter = flt@filter$z
tskewfilter = flt@filter$tskew
tshape1filter = flt@filter$tshape1
tshape2filter =flt@filter$tshape2
tempskewfilter = flt@filter$tempskew
tempshape1filter = flt@filter$tempshape1
tempshape2filter = flt@filter$tempshape2
seriesfor = sigmafor = matrix(NA, nrow = n.roll+1, ncol = n.ahead)
skewnessFor = kurtosisFor = matrix(NA,nrow = n.roll + 1, ncol = n.ahead)
# n.roll x n.ahead (n.ahead=1 generted by model)
# n.ahead>1 by simulation
rownames(seriesfor) = rownames(sigmafor) = rownames(skewnessFor) = rownames(kurtosisFor)= as.character(index[N:(N+n.roll)])
colnames(seriesfor) = colnames(sigmafor) = colnames(skewnessFor) = colnames(kurtosisFor) = paste("T+", 1:n.ahead, sep="")
mx = model$maxOrder #At the moment, the package only allows for maximum lag of 1 in both mean, sigma and higher moment equations.
# Hence mx is always 1
#constmean = unname(coefacd(fit)["mu"])
for(i in 1:(n.roll+1)){
spec = fspec
presig = tail(sigmafilter[1:(N+i-1)], mx)
preskew = tail(tskewfilter[1:(N+i-1)], mx)
preshape1 = tail(tshape1filter[1:(N+i-1)], mx)
preshape2 = tail(tshape2filter[1:(N+i-1)], mx)
prereturns = tail(data[1:(N+i-1)], mx)
sim = acdpath(spec, n.sim = horizon, n.start = 0, m.sim = m.sim,
presigma = presig, preskew = preskew,
preshape1 = preshape1,preshape2 = preshape2, prereturns = prereturns,
preresiduals = NA, rseed = rseed,
cluster = cluster)
seriesfor[i, ] = mean(colSums(sim@path$seriesSim))
sigmafor[i, ] = sqrt(mean(colSums(sim@path$sigmaSim^2)))
skewnessFor[i, ] = PerformanceAnalytics::skewness(colSums(sim@path$seriesSim))
kurtosisFor[i, ] = PerformanceAnalytics::kurtosis(colSums(sim@path$seriesSim),method = "excess")
}
} else{
seriesfor = sigmafor = matrix(NA, nrow = 1, ncol = n.ahead)
skewnessFor = kurtosisFor = matrix(NA,nrow = 1, ncol = n.ahead)
# n.roll x n.ahead (n.ahead=1 generted by model)
# n.ahead>1 by simulation
rownames(seriesfor) = rownames(sigmafor) = rownames(skewnessFor) = rownames(kurtosisFor)= as.character(index[N])
colnames(seriesfor) = colnames(sigmafor) = colnames(skewnessFor) = colnames(kurtosisFor) = paste("T+", 1:n.ahead, sep="")
mx = model$maxOrder
presig = as.numeric(tail(sigmaAcd(fit),mx))
preskew = as.numeric(tail(skew(fit),mx))
preshape1 = as.numeric(tail(shape1(fit),mx))
preshape2 = as.numeric(tail(shape2(fit),mx))
preresiduals = as.numeric(tail(residualsAcd(fit),mx))
prereturns = as.numeric(tail(fit@model$modeldata$data[N],mx))
sim = acdsim(fit, n.sim = horizon, m.sim = m.sim,
presigma = presig, preskew = preskew,
preshape1 = preshape1,preshape2 = preshape2, prereturns = prereturns,
preresiduals = preresiduals, rseed = rseed,
cluster = cluster)
seriesfor[,1] = mean(colSums(sim@simulation$seriesSim))
sigmafor[,1] = sqrt(mean(colSums(sim@simulation$sigmaSim^2)))
skewnessFor[,1] = PerformanceAnalytics::skewness(colSums(sim@simulation$seriesSim))
}
fcst = list()
fcst$n.ahead = n.ahead
fcst$n.roll = n.roll
fcst$N = N+ns
fcst$n.start = ns
fcst$seriesFor = seriesfor
fcst$sigmaFor = sigmafor
fcst$skewnessFor = skewnessFor
fcst$kurtosisFor = kurtosisFor
if(n.roll>0){
model$modeldata$sigma = flt@filter$sigma
model$modeldata$residuals = flt@filter$residuals
model$modeldata$tskew = flt@filter$tskew
model$modeldata$tshape1 = flt@filter$tshape1
model$modeldata$tshape2 = flt@filter$tshape2
model$modeldata$Pskewness = flt@filter$Pskewness
} else{
model = fit@model
}
ans = new("ACDforecast",
forecast = fcst,
model = model)
return(ans)
}
# CONTINUE HERE:
.acdforecast2L = function(spec, data = NULL, n.ahead = 10, n.roll = 0, out.sample = 0,
m.sim = 1000, cluster = NULL, skew0 = NULL,
shape0 = NULL, ...)
{
vmodel = spec@model$vmodel$model
xdata = rugarch:::.extractdata(data)
data = xdata$data
index = xdata$index
period = xdata$period
Nor = length(as.numeric(data))
ns = out.sample
N = Nor - ns
model = spec@model
modelinc = model$modelinc
idx = model$pidx
ipars = model$pars
pars = unlist(model$fixed.pars)
parnames = names(pars)
modelnames = rugarch:::.checkallfixed(spec)
if(is.na(all(match(modelnames, parnames), 1:length(modelnames)))) {
cat("\nacdfilter-->error: parameters names do not match specification\n")
cat("Expected Parameters are: ")
cat(paste(modelnames))
cat("\n")
stop("Exiting", call. = FALSE)
}
fspec = spec
setfixed(fspec)<-as.list(pars)
if( n.roll > ns ) stop("\nacdforecast-->error: n.roll must not be greater than out.sample!")
# filter data (check external regressor data - must equal length of origData)
fcreq = ifelse(ns >= (n.ahead+n.roll), n.ahead+n.roll, ns)
# Generate the 1 extra ahead forecast
if((n.ahead+n.roll)<=ns){
tmp = xts(data[1:(N + fcreq)], index[1:(N + fcreq)])
} else{
tmp = xts(c(data[1:(N + fcreq)],0), c(index[1:(N + fcreq)], index[(N + fcreq)]+1))
}
flt = acdfilter(spec = fspec, data = tmp, n.old = N, skew0 = skew0, shape0 = shape0)
sigmafilter = flt@filter$sigma
resfilter = flt@filter$residuals
zfilter = flt@filter$z
tskewfilter = flt@filter$tskew
tshapefilter = flt@filter$tshape
tempskewfilter = flt@filter$tempskew
tempshapefilter = flt@filter$tempshape
seriesfor = sigmafor = matrix(NA, nrow = n.roll+1, ncol = n.ahead)
skewnessFor = kurtosisFor = matrix(NA,nrow = n.roll + 1, ncol = n.ahead)
# n.roll x n.ahead (n.ahead=1 generted by model)
# n.ahead>1 by simulation
rownames(seriesfor) = rownames(sigmafor) = rownames(skewnessFor) = rownames(kurtosisFor)= as.character(index[N:(N+n.roll)])
colnames(seriesfor) = colnames(sigmafor) = colnames(skewnessFor) = colnames(kurtosisFor) = paste("T+", 1:n.ahead, sep="")
mx = model$maxOrder #At the moment, the package only allows for maximum lag of 1 in both mean, sigma and higher moment equations.
# Hence mx is always 1
#constmean = unname(coefacd(fit)["mu"])
for(i in 1:(n.roll+1)){
spec = fspec
presig = tail(sigmafilter[1:(N+i-1)], mx)
preskew = tail(tskewfilter[1:(N+i-1)], mx)
preshape1 = tail(tshape1filter[1:(N+i-1)], mx)
preshape2 = tail(tshape2filter[1:(N+i-1)], mx)
prereturns = tail(data[1:(N+i-1)], mx)
sim = acdpath(spec, n.sim = horizon, n.start = 0, m.sim = m.sim,
presigma = presig, preskew = preskew,
preshape1 = preshape1,preshape2 = preshape2, prereturns = prereturns,
preresiduals = NA, rseed = rseed,
cluster = cluster)
seriesfor[i, ] = mean(colSums(sim@path$seriesSim))
sigmafor[i, ] = sqrt(mean(colSums(sim@path$sigmaSim^2)))
skewnessFor[i, ] = PerformanceAnalytics::skewness(colSums(sim@path$seriesSim))
kurtosisFor[i, ] = PerformanceAnalytics::kurtosis(colSums(sim@path$seriesSim),method = "excess")
}
fcst = list()
fcst$n.ahead = n.ahead
fcst$n.roll = n.roll
fcst$N = N+ns
fcst$n.start = ns
fcst$seriesFor = seriesfor
fcst$sigmaFor = sigmafor
fcst$skewnessFor = skewnessfor
fcst$kurtosisFor = kurtosisFor
model$modeldata$sigma = flt@filter$sigma
model$modeldata$residuals = flt@filter$residuals
model$modeldata$tskew = flt@filter$tskew
model$modeldata$tshape = flt@filter$tshape
ans = new("ACDforecast",
forecast = fcst,
model = model)
return(ans)
}
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.