Nothing
R/BreakAnnualAggLossDevModelInput.R
In BALD: Robust Loss Development Using MCMC
Defines functions
makeBreakAnnualInput
Documented in
makeBreakAnnualInput
##################################################################################################
## ##
## BALD is an R-package. ##
## It is a Bayesian time series model of loss development. ##
## Features include skewed Student-t distribution with time-varying scale parameters, ##
## an expert prior for the calendar year effect, ##
## and accommodation for structural breaks in the consumption path of development years. ##
## It is an update for the older package lossDev as it has been stopped supported. ##
## ##
## Copyright (c) 2018 Frank A. Schmid, ##
## ##
## This file is part of BALD. ##
## ##
## lossDev is free software: you can redistribute it and/or modify ##
## it under the terms of the GNU General Public License as published by ##
## the Free Software Foundation, either version 3 of the License, or ##
## (at your option) any later version. ##
## ##
## This program is distributed in the hope that it will be useful, ##
## but WITHOUT ANY WARRANTY; without even the implied warranty of ##
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ##
## GNU General Public License for more details. ##
## ##
## You should have received a copy of the GNU General Public License ##
## along with this program. If not, see <https://www.gnu.org/licenses/>. ##
## ##
##################################################################################################
##This is intended to be a final class for models with a break.
##This class is derived from the AnnualAggLosDevModelInput class.
##' @include zzz.R
##' @include AnnualAggLossDevModelInput.R
NULL
##' The final input class for models with a break.
##'
##' \code{BreakAnnualAggLossDevModelInput} is the final input class for models with a break.
##' @name BreakAnnualAggLossDevModelInput-class
##' @docType class
##' @seealso \code{\linkS4class{LossDevModelInput}}
##' \code{\linkS4class{StandardAnnualAggLossDevModelInput}}
setClass(
'BreakAnnualAggLossDevModelInput',
representation(
rangeForFirstYearInNewRegime='integer', #The range of exposure years describing the break.
priorsForFirstYearInNewRegime='numeric', #The prior for the break.
priorForKnotPositionsPreBreak='numeric',
priorForKnotPositionsPostBreak='numeric',
priorForNumberOfKnotsPreBreak='numeric',
priorForNumberOfKnotsPostBreak='numeric'),
contains='AnnualAggLossDevModelInput')
##' Create an Object of class \code{BreakAnnualAggLossDevModelInput}.
##'
##' The loss development models require a lot of input. Much of the input is directly dependent on the values of other input.
##' As such, this function facilitates much of the work of setting model parameters and determining which output to collect.
##' See \code{vignette('BALD')}.
##'
##' The function creates an object of class \code{BreakAnnualAggLossDevModelInput}.
##'
##' Many arguments and much functionality is described in \code{\link{makeStandardAnnualInput}}.
##'
##' \describe{
##' \item{\code{first.year.in.new.regime}}{
##' The break model allows for a structural break along the exposure year axis in the consumption path.
##' The slope of the pre-break consumption path is used to extend the post-break consumption path.
##' The time when the break occurs can either be specified with 100% certainty by the user, or it can be estimated by the model.
##' To specify \emph{the} first exposure year in which a new consumption path applies, the user should supply a single value to the parameter \code{first.year.in.new.regime}.
##' To have the model estimate the point in time when the break occurs, the user should supply a range (min and max) for the possible values to \code{first.year.in.new.regime}.
##' Note also that the number of rows above and below the break must be at least 4 and as such a triangle of size 8 is the smallest triangle which can be estimated with a break.
##' }
##' \item{\code{prior.for.first.year.in.new.regime}}{
##' The prior for the \code{first.year.in.new.regime} is a (scaled and discretized) beta distribtion.
##' A value of \code{c(1,1)} would indicate a uniform distribution.
##' See \code{\link{firstYearInNewRegime}}.
##' }
##' \item{\code{prior.for.knot.locations.pre.break} and \code{prior.for.knot.locations.post.break}}{
##' If these values are \code{NA} (the default), then \code{prior.for.knot.locations.pre.break} will be assigned a value of 2.
##' And \code{prior.for.knot.locations.post.break} wil be assigned a value of \code{1 + (num.years.in.post.break.period + 0.5 * num.years.in.break.period)/num.years.in.triangle.}.
##' These values must either both be \code{NA} or both be set to numbers.
##' }
##' }
##'
##' @param incremental.payments A square matrix of incremental payments. Row names should correspond to the exposure year. Only upper-left (including the diagonal) of Triangle may have non-missing values. Lower-right must be \code{NA}.
##' @param first.year.in.new.regime May be one of two types. 1) A single numeric value. 2) A numeric vector of length 2. See Details.
##' @param prior.for.first.year.in.new.regime A numeric vector of length 2. See Details.
##'
##' @param extra.exp.years A single integer value (\code{total.exp.years} overrides) greater than or equal to 1 (default is 1) specifying the number of additional exposure years (or rows in the triangle) to project.
##' @param extra.dev.years A single integer value (\code{total.dev.years} overrides) greater than or equal to 1 (default is 1) specifying the additional number of development years (or columns in the triangle) to project.
##'
##' @param non.stoch.inflation.rate May be one of three types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) A vector of numerics (of specific length). 3) A matrix of numerics (of specific dim).
##' @param non.stoch.inflation.weight May be one of three types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) A vector of numerics (of specific length). 3) A matrix of numerics (of specific dim).
##'
##' @param stoch.inflation.rate May be one of two types (See \emph{Inflation Rate} in Details). 1) A single numeric value of \emph{zero}. 2) A vector of numerics (of specific length).
##' @param stoch.inflation.weight May be one of three types (See \emph{Inflation Rate} in Details). 1) A single numerical value. 2) A vector of numerics (of specific length). 3) A matrix of numerics (of specific dim).
##' @param stoch.inflation.lower.bound May be one of three types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) A vector of numerics (of specific length). 3) A matrix of numerics (of specific dim).
##' @param stoch.inflation.upper.bound May be one of three types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) A vector of numerics (of specific length). 3) A matrix of numerics (of specific dim).
##' @param known.stoch.inflation.mean May be one of two types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) \code{NA}.
##' @param known.stoch.inflation.persistence May be one of two types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) \code{NA}.
##'
##' @param total.exp.years A single integer value (overrides \code{extra.exp.years}) specifying the last exposure year to project. Must be at least the number of rows in \code{incremental.payments} + 1.
##' @param total.dev.years A single integer value (overrides \code{extra.dev.years}) specifying the last development year to project. Must be at least the number of columns in \code{incremental.payments} + 1 .
##'
##' @param cumulative.payments A numeric matrix with the same dim and dimnames as \code{incremental.payments}. Must be a possible cumulative payment triangle of \code{incremental.payments}. (See \emph{Cumulative Payments} Section).
##'
##' @param exp.year.type A single character value indicating the type of exposure years: \sQuote{ambiguous}, \sQuote{py}, and \sQuote{ay} mean \sQuote{Exposure Year}, \sQuote{Policy Year}, and \sQuote{Accident Year}; respectively.
##' @param prior.for.knot.locations.pre.break A single numeric value of at least 1. The prior for the location of knots is a scaled beta with parameters \code{c(1,prior.for.knot.locations.pre.break)}.
##' @param prior.for.number.of.knots.pre.break A two element vector giving the paramters for the prior number of knots.
##' @param prior.for.knot.locations.post.break See \code{prior.for.knot.locations.pre.break}. Large values produce stable consumption paths at high development years.
##' @param prior.for.number.of.knots.post.break A two element vector giving the paramters for the prior number of knots.
##'
##' @param use.skew.t A logical value. If \code{TRUE} the model assumes the observed and estimated log incremental payments are realizations from a skewed \eqn{t} distribution; if \code{FALSE} it will assume zero skewness. (See Reference.)
##' @param bound.for.skewness.parameter A positive numerical value representing the symetric boundaries for the skewness parameter. In most cases, the default should be sufficient. Ignored if \code{use.skew.t=FALSE}.
##' @param last.column.with.scale.innovation A single integer must be at least 1 and at most the number of columns in \code{incremental.payments}. See \emph{Measurment Error-Second Order Random Walk} in Details.
##' @param use.ar1.in.calendar.year A logical value. The calendar year effect errors may (at the users digression) include an autoregressive process of order 1. \code{TRUE} turns on the ar1 process, \code{FALSE} turns it off.
##' @param use.ar1.in.exposure.growth A logical value. The exposure growth errors may (at the users discretion) include an autoregressive process of order 1. \code{TRUE} (the Default) turns on the ar1 process, \code{FALSE} turns it off.
##'
##' @param projected.rate.of.decay May be one of three types (See \emph{Projected Rate of Decay} in Details). 1) \code{NA}. 2) A matrix of numerics (of specific dim). 3) A named list.
##'
##'
##' @references
##' Kim, Y., and J. McCulloch (2007) \dQuote{The Skew-Student Distribution with Application to U.S. Stock Market Returns and the Equity Premium,} Department of Economics, Ohio State University, October 2007
##'
##'
##' @return An object of class \code{AggModelInput}. This the model specified by the returned object must then be estimated using the function \code{runLossDevModel}.
##' @export
##' @examples
##' rm(list=ls())
##' library(BALD)
##' data(CumulativeAutoBodilyInjuryTriangle)
##' CumulativeAutoBodilyInjuryTriangle <- as.matrix(CumulativeAutoBodilyInjuryTriangle)
##' sample.col <- (dim(CumulativeAutoBodilyInjuryTriangle)[2] - 6:0)
##' print(decumulate(CumulativeAutoBodilyInjuryTriangle)[1:7, sample.col])
##' data(HPCE)
##' HPCE <- as.matrix(HPCE)[,1]
##' HPCE.rate <- HPCE[-1] / HPCE[-length(HPCE)] - 1
##' print(HPCE.rate[(-10):0 + length(HPCE.rate)])
##' HPCE.years <- as.integer(names(HPCE.rate))
##' max.exp.year <- max(as.integer(
##' dimnames(CumulativeAutoBodilyInjuryTriangle)[[1]]))
##' years.to.keep <- HPCE.years <= max.exp.year + 3
##' HPCE.rate <- HPCE.rate[years.to.keep]
##' break.model.input <- makeBreakAnnualInput(
##' cumulative.payments = CumulativeAutoBodilyInjuryTriangle,
##' stoch.inflation.weight = 1,
##' non.stoch.inflation.weight = 0,
##' stoch.inflation.rate = HPCE.rate,
##' first.year.in.new.regime = c(1986, 1987),
##' prior.for.first.year.in.new.regime=c(2,1),
##' exp.year.type = 'ay',
##' extra.dev.years = 5,
##' use.skew.t = TRUE,
##' bound.for.skewness.parameter=5)
##'
##' @usage
##' makeBreakAnnualInput(
##' incremental.payments=decumulate(cumulative.payments),
##' first.year.in.new.regime=trunc(median(as.integer(dimnames(incremental.payments)[[1]]))),
##' prior.for.first.year.in.new.regime=c(2,2),
##' extra.dev.years=1,
##' extra.exp.years=1,
##' non.stoch.inflation.rate=0,
##' non.stoch.inflation.weight=1,
##' stoch.inflation.rate=0,
##' stoch.inflation.weight=1-non.stoch.inflation.weight,
##' stoch.inflation.lower.bound=-1,
##' stoch.inflation.upper.bound=Inf,
##' known.stoch.inflation.mean=NA,
##' known.stoch.inflation.persistence=NA,
##' total.dev.years=extra.dev.years+dim(incremental.payments)[2],
##' total.exp.years=extra.exp.years+dim(incremental.payments)[1],
##' cumulative.payments=cumulate(incremental.payments),
##' exp.year.type=c('ambiguous', 'py', 'ay'),
##' prior.for.knot.locations.pre.break=NA,
##' prior.for.number.of.knots.pre.break=c(3, 1/7),
##' prior.for.knot.locations.post.break=NA,
##' prior.for.number.of.knots.post.break=c(3, 1/7),
##' use.skew.t=FALSE,
##' bound.for.skewness.parameter=10,
##' last.column.with.scale.innovation=dim(incremental.payments)[2],
##' use.ar1.in.calendar.year=FALSE,
##' use.ar1.in.exposure.growth=TRUE,
##' projected.rate.of.decay=NA)
makeBreakAnnualInput <- function(incremental.payments=decumulate(cumulative.payments),
first.year.in.new.regime=trunc(median(as.integer(dimnames(incremental.payments)[[1]]))),
prior.for.first.year.in.new.regime=c(2,2),
extra.dev.years=1,
extra.exp.years=1,
non.stoch.inflation.rate=0,
non.stoch.inflation.weight=1,
stoch.inflation.rate=0,
stoch.inflation.weight=1-non.stoch.inflation.weight,
stoch.inflation.lower.bound=-1,
stoch.inflation.upper.bound=Inf,
known.stoch.inflation.mean=NA,
known.stoch.inflation.persistence=NA,
total.dev.years=extra.dev.years+dim(incremental.payments)[2],
total.exp.years=extra.exp.years+dim(incremental.payments)[1],
cumulative.payments=cumulate(incremental.payments),
exp.year.type=c('ambiguous', 'py', 'ay'),
prior.for.knot.locations.pre.break=NA,
prior.for.number.of.knots.pre.break=c(3,1/7),
prior.for.knot.locations.post.break=NA,
prior.for.number.of.knots.post.break=c(3,1/7),
use.skew.t=FALSE,
bound.for.skewness.parameter=10,
last.column.with.scale.innovation=dim(incremental.payments)[2],
use.ar1.in.calendar.year=FALSE,
use.ar1.in.exposure.growth=TRUE,
projected.rate.of.decay=NA)
{
standard.ans <- makeStandardAnnualInput(incremental.payments=incremental.payments,
extra.dev.years=extra.dev.years,
extra.exp.years=extra.exp.years,
non.stoch.inflation.rate=non.stoch.inflation.rate,
non.stoch.inflation.weight=non.stoch.inflation.weight,
stoch.inflation.rate=stoch.inflation.rate,
stoch.inflation.weight=stoch.inflation.weight,
stoch.inflation.lower.bound=stoch.inflation.lower.bound,
stoch.inflation.upper.bound=stoch.inflation.upper.bound,
known.stoch.inflation.mean=known.stoch.inflation.mean,
known.stoch.inflation.persistence=known.stoch.inflation.persistence,
total.dev.years=total.dev.years,
total.exp.years=total.exp.years,
cumulative.payments=cumulative.payments,
exp.year.type=exp.year.type,
use.skew.t=use.skew.t,
bound.for.skewness.parameter=bound.for.skewness.parameter,
last.column.with.scale.innovation=last.column.with.scale.innovation,
use.ar1.in.calendar.year=use.ar1.in.calendar.year,
use.ar1.in.exposure.growth=use.ar1.in.exposure.growth,
projected.rate.of.decay=projected.rate.of.decay)
ans <- new('BreakAnnualAggLossDevModelInput')
for(s in slotNames(standard.ans))
{
if(s %in% slotNames(ans))
{
slot(ans, s) <- slot(standard.ans, s)
}
}
rm(standard.ans)
##perform some constancy checks
if(getTriDim(ans)[1] < 8)
stop('The triangle supplied to "incremental.payments" is too small to run with a break. Try running the standard model.')
suppressWarnings(int.first.year.in.new.regime <- as.integer(first.year.in.new.regime))
if(any(int.first.year.in.new.regime != first.year.in.new.regime))
stop('"first.year.in.new.regime" must be coercible to an integer')
if(length(int.first.year.in.new.regime) > 2 || length(int.first.year.in.new.regime) == 0 || !all(int.first.year.in.new.regime %in% ans@exposureYears))
stop('"first.year.in.new.regime" must have a length of either 1 or 2 and must only contain values in the exposure years of incremental.payments."')
if(length(int.first.year.in.new.regime) == 1)
{
ans@rangeForFirstYearInNewRegime <- rep(int.first.year.in.new.regime, 2)
} else {
if(int.first.year.in.new.regime[1] > int.first.year.in.new.regime[2])
{
warning('"int.first.year.in.new.regime[1] > int.first.year.in.new.regime[2]" these values will be resorted')
int.first.year.in.new.regime <- sort(int.first.year.in.new.regime)
}
ans@rangeForFirstYearInNewRegime <- int.first.year.in.new.regime
}
if(length(prior.for.knot.locations.pre.break) == 1 && is.na(prior.for.knot.locations.pre.break) &&
length(prior.for.knot.locations.post.break) == 1 && is.na(prior.for.knot.locations.post.break))
{
ans@priorForKnotPositionsPreBreak <- 2
ans@priorForKnotPositionsPostBreak <- 1 +
(sum(ans@exposureYears > max(ans@rangeForFirstYearInNewRegime)) + .5 * (1 + diff(ans@rangeForFirstYearInNewRegime))) /
length(ans@exposureYears)
} else {
if(!is.numeric(prior.for.knot.locations.pre.break) || length(prior.for.knot.locations.pre.break) != 1)
stop('"prior.for.knot.locations.pre.break" and "prior.for.knot.locations.post.break" must both be numeric of length 1.')
if(prior.for.knot.locations.pre.break < 1)
stop('"prior.for.knot.locations.pre.break" must be at least 1')
ans@priorForKnotPositionsPreBreak <- prior.for.knot.locations.pre.break
if(!is.numeric(prior.for.knot.locations.post.break) || length(prior.for.knot.locations.post.break) != 1)
stop('"prior.for.knot.locations.pre.break" and "prior.for.knot.locations.post.break" must both be numeric of length 1')
if(prior.for.knot.locations.post.break < 1)
stop('"prior.for.knot.locations.post.break" must be at least 1')
ans@priorForKnotPositionsPostBreak <- prior.for.knot.locations.post.break
}
if(!is.numeric(prior.for.number.of.knots.pre.break) || length(prior.for.number.of.knots.pre.break) != 2)
stop('"prior.for.number.of.knots.pre.break" must be a numeric of length 2')
if(prior.for.number.of.knots.pre.break[1] <= 0)
stop('"prior.for.number.of.knots.pre.break[1]" must be greater than zero')
if(prior.for.number.of.knots.pre.break[2] <= 0 ||prior.for.number.of.knots.pre.break[2] >= 1 )
stop('"prior.for.number.of.knots.pre.break[2]" must be greater than zero but less than one')
ans@priorForNumberOfKnotsPreBreak <- prior.for.number.of.knots.pre.break
if(!is.numeric(prior.for.number.of.knots.post.break) || length(prior.for.number.of.knots.post.break) != 2)
stop('"prior.for.number.of.knots.post.break" must be a numeric of length 2')
if(prior.for.number.of.knots.post.break[1] <= 0)
stop('"prior.for.number.of.knots.post.break[1]" must be greater than zero')
if(prior.for.number.of.knots.post.break[2] <= 0 ||prior.for.number.of.knots.post.break[2] >= 1 )
stop('"prior.for.number.of.knots.post.break[2]" must be greater than zero but less than one')
ans@priorForNumberOfKnotsPostBreak <- prior.for.number.of.knots.post.break
if(sum(ans@exposureYears < ans@rangeForFirstYearInNewRegime[1]) < 4)
stop('The minimum value for "first.year.in.new.regime" is too small. There must be at least 4 years in the pre-break period.')
if(sum(ans@exposureYears >= ans@rangeForFirstYearInNewRegime[2]) < 4)
stop('The maximum value for "first.year.in.new.regime" is too large. There must be at least 4 years in the post-break period.')
if(!is.numeric(prior.for.first.year.in.new.regime) || length(prior.for.first.year.in.new.regime) != 2)
stop('The value supplied for "prior.for.first.year.in.new.regime" must be a numeric vector of length 2.')
if(any(prior.for.first.year.in.new.regime <=0))
stop('"prior.for.first.year.in.new.regime" must be greater than 0')
ans@priorsForFirstYearInNewRegime <- prior.for.first.year.in.new.regime
ans@outputType <- 'BreakAnnualAggLossDevModelOutput'
ans@modelFile <- 'break.model.txt'
return(invisible(ans))
}
##' A method to collect all the needed model input specific to the break model.
##'
##' There are currently two types of \code{AnnualAggLossDevModel}s (break and standard). These models have many data elements in common.
##' This method appends only the elements specific to the break model onto the list created by a call to \code{NextMethod()}.
##'
##' The following elements are appended:
##' \describe{
##' \item{\code{x.0}}{Two values. The lower bound for the number of knots. First is for first spline.}
##' \item{\code{x.r}}{Two values. The upper bound for the number of knots. First is for first spline.}
##' \item{\code{break.row}}{Two integer values given the (inclusive) range of rows in which the first year in the new regime could occur.}
##' \item{\code{break.row.priors}}{The parameters for the beta distribution which serves as the prior for the location of the structural break.}
##' \item{\code{K.trim}}{The maximum number of columns in the post-break triangle.}
##' \item{\code{beta.prior}}{A matrix giving the prior for the location of knots. First column is for the pre-break spline. Second is for the post-break spline.}
##' \item{\code{mu.number.of.knots.prior}}{A matrix giving the prior for the mean of the number of knots. First column is for the pre-break spline. Second is for the post-break spline}
##' \item{\code{number.of.knots.ubound}}{A vector giving the upper bound for the number of knots. First is for the pre-break spline. Second is for the post-break spline}
##' }
##' @name getJagsData,BreakAnnualLossDevModelInput-method
##' @param object An object of type \code{BreakAnnualAggLossDevModelInput} from which to collect the needed model input.
##' @return A named list of the specific model elements. See details for more info.
##' @docType methods
setMethod(
'getJagsData',
signature(object='BreakAnnualAggLossDevModelInput'),
function(object)
{
ans <- callNextMethod()
##return(ans)
log.inc.index <- ans$log.inc.index
##calculate the last column with two non-missing values
x.r <- function()
{
log.inc <- ans$log.inc
for(i in dim(log.inc)[2]:1)
{
tmp <- sum(!is.na(log.inc[,i]))
if(tmp >= 2)
return(i)
}
stop('"log.inc" does not have a column with at least two non missing values')
}
x.r.post.break <- function()
{
log.inc <- ans$log.inc
log.inc <- log.inc[object@exposureYears >= max(object@rangeForFirstYearInNewRegime),]
for(i in dim(log.inc)[2]:1)
{
tmp <- sum(!is.na(log.inc[,i]))
if(tmp >= 2)
return(i)
}
stop('"log.inc" for the post period does not have a column with at least two non missing values')
}
ans$x.0 <- c(1, 1)
ans$x.r <- c(x.r(), x.r.post.break())
ans$break.row <- match(object@rangeForFirstYearInNewRegime, object@exposureYears)
ans$break.row.priors <- object@priorsForFirstYearInNewRegime
ans$K.trim <- ans$K - ans$break.row[1] + 1
ans$beta.prior <- array(NA, c(2,2))
ans$beta.prior[,1] <- c(1,object@priorForKnotPositionsPreBreak)
ans$beta.prior[,2] <- c(1,object@priorForKnotPositionsPostBreak)
ans$mu.number.of.knots.prior <- array(NA, c(2, 2))
ans$mu.number.of.knots.prior[,1] <- object@priorForNumberOfKnotsPreBreak
ans$mu.number.of.knots.prior[2,1] <- (1 - ans$mu.number.of.knots.prior[2,1]) / ans$mu.number.of.knots.prior[2,1]
ans$mu.number.of.knots.prior[,2] <- object@priorForNumberOfKnotsPostBreak
ans$mu.number.of.knots.prior[2,2] <- (1 - ans$mu.number.of.knots.prior[2,2]) / ans$mu.number.of.knots.prior[2,2]
ans$number.of.knots.ubound <- numeric(2)
ans$number.of.knots.ubound[1] <- trunc(ans$K/2) + 1
ans$number.of.knots.ubound[2] <- trunc(ans$K.trim/2) + 1
return(ans)
})
##' A method to collect all the needed model initial values unique to the break model.
##'
##' There are currently two types of \code{AnnualAggLossDevModel}s (break and standard). Code needed to create initial values specific the break model is placed in this method.
##' This method returns a parameterless function which when called first calls the function returned by \code{NextMethod()} and then to the list returned by that function appends the following initial values.
##'
##' \describe{
##' \item{\code{R.}}{The initial values for the spline node. Needed because dspline cannot create initial values.}
##' }
##' @name getJagsInits,BreakAnnualLossDevModelInput-method
##' @param object An object of type \code{BreakAnnualAggLossDevModelInput} from which to collect the needed model initial values.
##' @return A named list of the specific model elements. See details for more info.
##' @docType methods
##' @seealso \code{\link{getJagsInits}}
setMethod(
'getJagsInits',
signature(object='BreakAnnualAggLossDevModelInput'),
function(object)
{
super.f <- callNextMethod()
K <- getTriDim(object)[1]
R. <- array(0, c(K, 6))
R.[,2] <- 1
function()
{
ans <- super.f()
ans$R. <- R.
return(ans)
}
})
Try the BALD package in your browser
Any scripts or data that you put into this service are public.
BALD documentation built on May 2, 2019, 6:51 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 makeBreakAnnualInput
Documented in makeBreakAnnualInput
##################################################################################################
## ##
## BALD is an R-package. ##
## It is a Bayesian time series model of loss development. ##
## Features include skewed Student-t distribution with time-varying scale parameters, ##
## an expert prior for the calendar year effect, ##
## and accommodation for structural breaks in the consumption path of development years. ##
## It is an update for the older package lossDev as it has been stopped supported. ##
## ##
## Copyright (c) 2018 Frank A. Schmid, ##
## ##
## This file is part of BALD. ##
## ##
## lossDev is free software: you can redistribute it and/or modify ##
## it under the terms of the GNU General Public License as published by ##
## the Free Software Foundation, either version 3 of the License, or ##
## (at your option) any later version. ##
## ##
## This program is distributed in the hope that it will be useful, ##
## but WITHOUT ANY WARRANTY; without even the implied warranty of ##
## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the ##
## GNU General Public License for more details. ##
## ##
## You should have received a copy of the GNU General Public License ##
## along with this program. If not, see <https://www.gnu.org/licenses/>. ##
## ##
##################################################################################################
##This is intended to be a final class for models with a break.
##This class is derived from the AnnualAggLosDevModelInput class.
##' @include zzz.R
##' @include AnnualAggLossDevModelInput.R
NULL
##' The final input class for models with a break.
##'
##' \code{BreakAnnualAggLossDevModelInput} is the final input class for models with a break.
##' @name BreakAnnualAggLossDevModelInput-class
##' @docType class
##' @seealso \code{\linkS4class{LossDevModelInput}}
##' \code{\linkS4class{StandardAnnualAggLossDevModelInput}}
setClass(
'BreakAnnualAggLossDevModelInput',
representation(
rangeForFirstYearInNewRegime='integer', #The range of exposure years describing the break.
priorsForFirstYearInNewRegime='numeric', #The prior for the break.
priorForKnotPositionsPreBreak='numeric',
priorForKnotPositionsPostBreak='numeric',
priorForNumberOfKnotsPreBreak='numeric',
priorForNumberOfKnotsPostBreak='numeric'),
contains='AnnualAggLossDevModelInput')
##' Create an Object of class \code{BreakAnnualAggLossDevModelInput}.
##'
##' The loss development models require a lot of input. Much of the input is directly dependent on the values of other input.
##' As such, this function facilitates much of the work of setting model parameters and determining which output to collect.
##' See \code{vignette('BALD')}.
##'
##' The function creates an object of class \code{BreakAnnualAggLossDevModelInput}.
##'
##' Many arguments and much functionality is described in \code{\link{makeStandardAnnualInput}}.
##'
##' \describe{
##' \item{\code{first.year.in.new.regime}}{
##' The break model allows for a structural break along the exposure year axis in the consumption path.
##' The slope of the pre-break consumption path is used to extend the post-break consumption path.
##' The time when the break occurs can either be specified with 100% certainty by the user, or it can be estimated by the model.
##' To specify \emph{the} first exposure year in which a new consumption path applies, the user should supply a single value to the parameter \code{first.year.in.new.regime}.
##' To have the model estimate the point in time when the break occurs, the user should supply a range (min and max) for the possible values to \code{first.year.in.new.regime}.
##' Note also that the number of rows above and below the break must be at least 4 and as such a triangle of size 8 is the smallest triangle which can be estimated with a break.
##' }
##' \item{\code{prior.for.first.year.in.new.regime}}{
##' The prior for the \code{first.year.in.new.regime} is a (scaled and discretized) beta distribtion.
##' A value of \code{c(1,1)} would indicate a uniform distribution.
##' See \code{\link{firstYearInNewRegime}}.
##' }
##' \item{\code{prior.for.knot.locations.pre.break} and \code{prior.for.knot.locations.post.break}}{
##' If these values are \code{NA} (the default), then \code{prior.for.knot.locations.pre.break} will be assigned a value of 2.
##' And \code{prior.for.knot.locations.post.break} wil be assigned a value of \code{1 + (num.years.in.post.break.period + 0.5 * num.years.in.break.period)/num.years.in.triangle.}.
##' These values must either both be \code{NA} or both be set to numbers.
##' }
##' }
##'
##' @param incremental.payments A square matrix of incremental payments. Row names should correspond to the exposure year. Only upper-left (including the diagonal) of Triangle may have non-missing values. Lower-right must be \code{NA}.
##' @param first.year.in.new.regime May be one of two types. 1) A single numeric value. 2) A numeric vector of length 2. See Details.
##' @param prior.for.first.year.in.new.regime A numeric vector of length 2. See Details.
##'
##' @param extra.exp.years A single integer value (\code{total.exp.years} overrides) greater than or equal to 1 (default is 1) specifying the number of additional exposure years (or rows in the triangle) to project.
##' @param extra.dev.years A single integer value (\code{total.dev.years} overrides) greater than or equal to 1 (default is 1) specifying the additional number of development years (or columns in the triangle) to project.
##'
##' @param non.stoch.inflation.rate May be one of three types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) A vector of numerics (of specific length). 3) A matrix of numerics (of specific dim).
##' @param non.stoch.inflation.weight May be one of three types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) A vector of numerics (of specific length). 3) A matrix of numerics (of specific dim).
##'
##' @param stoch.inflation.rate May be one of two types (See \emph{Inflation Rate} in Details). 1) A single numeric value of \emph{zero}. 2) A vector of numerics (of specific length).
##' @param stoch.inflation.weight May be one of three types (See \emph{Inflation Rate} in Details). 1) A single numerical value. 2) A vector of numerics (of specific length). 3) A matrix of numerics (of specific dim).
##' @param stoch.inflation.lower.bound May be one of three types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) A vector of numerics (of specific length). 3) A matrix of numerics (of specific dim).
##' @param stoch.inflation.upper.bound May be one of three types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) A vector of numerics (of specific length). 3) A matrix of numerics (of specific dim).
##' @param known.stoch.inflation.mean May be one of two types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) \code{NA}.
##' @param known.stoch.inflation.persistence May be one of two types (See \emph{Inflation Rate} in Details). 1) A single numeric value. 2) \code{NA}.
##'
##' @param total.exp.years A single integer value (overrides \code{extra.exp.years}) specifying the last exposure year to project. Must be at least the number of rows in \code{incremental.payments} + 1.
##' @param total.dev.years A single integer value (overrides \code{extra.dev.years}) specifying the last development year to project. Must be at least the number of columns in \code{incremental.payments} + 1 .
##'
##' @param cumulative.payments A numeric matrix with the same dim and dimnames as \code{incremental.payments}. Must be a possible cumulative payment triangle of \code{incremental.payments}. (See \emph{Cumulative Payments} Section).
##'
##' @param exp.year.type A single character value indicating the type of exposure years: \sQuote{ambiguous}, \sQuote{py}, and \sQuote{ay} mean \sQuote{Exposure Year}, \sQuote{Policy Year}, and \sQuote{Accident Year}; respectively.
##' @param prior.for.knot.locations.pre.break A single numeric value of at least 1. The prior for the location of knots is a scaled beta with parameters \code{c(1,prior.for.knot.locations.pre.break)}.
##' @param prior.for.number.of.knots.pre.break A two element vector giving the paramters for the prior number of knots.
##' @param prior.for.knot.locations.post.break See \code{prior.for.knot.locations.pre.break}. Large values produce stable consumption paths at high development years.
##' @param prior.for.number.of.knots.post.break A two element vector giving the paramters for the prior number of knots.
##'
##' @param use.skew.t A logical value. If \code{TRUE} the model assumes the observed and estimated log incremental payments are realizations from a skewed \eqn{t} distribution; if \code{FALSE} it will assume zero skewness. (See Reference.)
##' @param bound.for.skewness.parameter A positive numerical value representing the symetric boundaries for the skewness parameter. In most cases, the default should be sufficient. Ignored if \code{use.skew.t=FALSE}.
##' @param last.column.with.scale.innovation A single integer must be at least 1 and at most the number of columns in \code{incremental.payments}. See \emph{Measurment Error-Second Order Random Walk} in Details.
##' @param use.ar1.in.calendar.year A logical value. The calendar year effect errors may (at the users digression) include an autoregressive process of order 1. \code{TRUE} turns on the ar1 process, \code{FALSE} turns it off.
##' @param use.ar1.in.exposure.growth A logical value. The exposure growth errors may (at the users discretion) include an autoregressive process of order 1. \code{TRUE} (the Default) turns on the ar1 process, \code{FALSE} turns it off.
##'
##' @param projected.rate.of.decay May be one of three types (See \emph{Projected Rate of Decay} in Details). 1) \code{NA}. 2) A matrix of numerics (of specific dim). 3) A named list.
##'
##'
##' @references
##' Kim, Y., and J. McCulloch (2007) \dQuote{The Skew-Student Distribution with Application to U.S. Stock Market Returns and the Equity Premium,} Department of Economics, Ohio State University, October 2007
##'
##'
##' @return An object of class \code{AggModelInput}. This the model specified by the returned object must then be estimated using the function \code{runLossDevModel}.
##' @export
##' @examples
##' rm(list=ls())
##' library(BALD)
##' data(CumulativeAutoBodilyInjuryTriangle)
##' CumulativeAutoBodilyInjuryTriangle <- as.matrix(CumulativeAutoBodilyInjuryTriangle)
##' sample.col <- (dim(CumulativeAutoBodilyInjuryTriangle)[2] - 6:0)
##' print(decumulate(CumulativeAutoBodilyInjuryTriangle)[1:7, sample.col])
##' data(HPCE)
##' HPCE <- as.matrix(HPCE)[,1]
##' HPCE.rate <- HPCE[-1] / HPCE[-length(HPCE)] - 1
##' print(HPCE.rate[(-10):0 + length(HPCE.rate)])
##' HPCE.years <- as.integer(names(HPCE.rate))
##' max.exp.year <- max(as.integer(
##' dimnames(CumulativeAutoBodilyInjuryTriangle)[[1]]))
##' years.to.keep <- HPCE.years <= max.exp.year + 3
##' HPCE.rate <- HPCE.rate[years.to.keep]
##' break.model.input <- makeBreakAnnualInput(
##' cumulative.payments = CumulativeAutoBodilyInjuryTriangle,
##' stoch.inflation.weight = 1,
##' non.stoch.inflation.weight = 0,
##' stoch.inflation.rate = HPCE.rate,
##' first.year.in.new.regime = c(1986, 1987),
##' prior.for.first.year.in.new.regime=c(2,1),
##' exp.year.type = 'ay',
##' extra.dev.years = 5,
##' use.skew.t = TRUE,
##' bound.for.skewness.parameter=5)
##'
##' @usage
##' makeBreakAnnualInput(
##' incremental.payments=decumulate(cumulative.payments),
##' first.year.in.new.regime=trunc(median(as.integer(dimnames(incremental.payments)[[1]]))),
##' prior.for.first.year.in.new.regime=c(2,2),
##' extra.dev.years=1,
##' extra.exp.years=1,
##' non.stoch.inflation.rate=0,
##' non.stoch.inflation.weight=1,
##' stoch.inflation.rate=0,
##' stoch.inflation.weight=1-non.stoch.inflation.weight,
##' stoch.inflation.lower.bound=-1,
##' stoch.inflation.upper.bound=Inf,
##' known.stoch.inflation.mean=NA,
##' known.stoch.inflation.persistence=NA,
##' total.dev.years=extra.dev.years+dim(incremental.payments)[2],
##' total.exp.years=extra.exp.years+dim(incremental.payments)[1],
##' cumulative.payments=cumulate(incremental.payments),
##' exp.year.type=c('ambiguous', 'py', 'ay'),
##' prior.for.knot.locations.pre.break=NA,
##' prior.for.number.of.knots.pre.break=c(3, 1/7),
##' prior.for.knot.locations.post.break=NA,
##' prior.for.number.of.knots.post.break=c(3, 1/7),
##' use.skew.t=FALSE,
##' bound.for.skewness.parameter=10,
##' last.column.with.scale.innovation=dim(incremental.payments)[2],
##' use.ar1.in.calendar.year=FALSE,
##' use.ar1.in.exposure.growth=TRUE,
##' projected.rate.of.decay=NA)
makeBreakAnnualInput <- function(incremental.payments=decumulate(cumulative.payments),
first.year.in.new.regime=trunc(median(as.integer(dimnames(incremental.payments)[[1]]))),
prior.for.first.year.in.new.regime=c(2,2),
extra.dev.years=1,
extra.exp.years=1,
non.stoch.inflation.rate=0,
non.stoch.inflation.weight=1,
stoch.inflation.rate=0,
stoch.inflation.weight=1-non.stoch.inflation.weight,
stoch.inflation.lower.bound=-1,
stoch.inflation.upper.bound=Inf,
known.stoch.inflation.mean=NA,
known.stoch.inflation.persistence=NA,
total.dev.years=extra.dev.years+dim(incremental.payments)[2],
total.exp.years=extra.exp.years+dim(incremental.payments)[1],
cumulative.payments=cumulate(incremental.payments),
exp.year.type=c('ambiguous', 'py', 'ay'),
prior.for.knot.locations.pre.break=NA,
prior.for.number.of.knots.pre.break=c(3,1/7),
prior.for.knot.locations.post.break=NA,
prior.for.number.of.knots.post.break=c(3,1/7),
use.skew.t=FALSE,
bound.for.skewness.parameter=10,
last.column.with.scale.innovation=dim(incremental.payments)[2],
use.ar1.in.calendar.year=FALSE,
use.ar1.in.exposure.growth=TRUE,
projected.rate.of.decay=NA)
{
standard.ans <- makeStandardAnnualInput(incremental.payments=incremental.payments,
extra.dev.years=extra.dev.years,
extra.exp.years=extra.exp.years,
non.stoch.inflation.rate=non.stoch.inflation.rate,
non.stoch.inflation.weight=non.stoch.inflation.weight,
stoch.inflation.rate=stoch.inflation.rate,
stoch.inflation.weight=stoch.inflation.weight,
stoch.inflation.lower.bound=stoch.inflation.lower.bound,
stoch.inflation.upper.bound=stoch.inflation.upper.bound,
known.stoch.inflation.mean=known.stoch.inflation.mean,
known.stoch.inflation.persistence=known.stoch.inflation.persistence,
total.dev.years=total.dev.years,
total.exp.years=total.exp.years,
cumulative.payments=cumulative.payments,
exp.year.type=exp.year.type,
use.skew.t=use.skew.t,
bound.for.skewness.parameter=bound.for.skewness.parameter,
last.column.with.scale.innovation=last.column.with.scale.innovation,
use.ar1.in.calendar.year=use.ar1.in.calendar.year,
use.ar1.in.exposure.growth=use.ar1.in.exposure.growth,
projected.rate.of.decay=projected.rate.of.decay)
ans <- new('BreakAnnualAggLossDevModelInput')
for(s in slotNames(standard.ans))
{
if(s %in% slotNames(ans))
{
slot(ans, s) <- slot(standard.ans, s)
}
}
rm(standard.ans)
##perform some constancy checks
if(getTriDim(ans)[1] < 8)
stop('The triangle supplied to "incremental.payments" is too small to run with a break. Try running the standard model.')
suppressWarnings(int.first.year.in.new.regime <- as.integer(first.year.in.new.regime))
if(any(int.first.year.in.new.regime != first.year.in.new.regime))
stop('"first.year.in.new.regime" must be coercible to an integer')
if(length(int.first.year.in.new.regime) > 2 || length(int.first.year.in.new.regime) == 0 || !all(int.first.year.in.new.regime %in% ans@exposureYears))
stop('"first.year.in.new.regime" must have a length of either 1 or 2 and must only contain values in the exposure years of incremental.payments."')
if(length(int.first.year.in.new.regime) == 1)
{
ans@rangeForFirstYearInNewRegime <- rep(int.first.year.in.new.regime, 2)
} else {
if(int.first.year.in.new.regime[1] > int.first.year.in.new.regime[2])
{
warning('"int.first.year.in.new.regime[1] > int.first.year.in.new.regime[2]" these values will be resorted')
int.first.year.in.new.regime <- sort(int.first.year.in.new.regime)
}
ans@rangeForFirstYearInNewRegime <- int.first.year.in.new.regime
}
if(length(prior.for.knot.locations.pre.break) == 1 && is.na(prior.for.knot.locations.pre.break) &&
length(prior.for.knot.locations.post.break) == 1 && is.na(prior.for.knot.locations.post.break))
{
ans@priorForKnotPositionsPreBreak <- 2
ans@priorForKnotPositionsPostBreak <- 1 +
(sum(ans@exposureYears > max(ans@rangeForFirstYearInNewRegime)) + .5 * (1 + diff(ans@rangeForFirstYearInNewRegime))) /
length(ans@exposureYears)
} else {
if(!is.numeric(prior.for.knot.locations.pre.break) || length(prior.for.knot.locations.pre.break) != 1)
stop('"prior.for.knot.locations.pre.break" and "prior.for.knot.locations.post.break" must both be numeric of length 1.')
if(prior.for.knot.locations.pre.break < 1)
stop('"prior.for.knot.locations.pre.break" must be at least 1')
ans@priorForKnotPositionsPreBreak <- prior.for.knot.locations.pre.break
if(!is.numeric(prior.for.knot.locations.post.break) || length(prior.for.knot.locations.post.break) != 1)
stop('"prior.for.knot.locations.pre.break" and "prior.for.knot.locations.post.break" must both be numeric of length 1')
if(prior.for.knot.locations.post.break < 1)
stop('"prior.for.knot.locations.post.break" must be at least 1')
ans@priorForKnotPositionsPostBreak <- prior.for.knot.locations.post.break
}
if(!is.numeric(prior.for.number.of.knots.pre.break) || length(prior.for.number.of.knots.pre.break) != 2)
stop('"prior.for.number.of.knots.pre.break" must be a numeric of length 2')
if(prior.for.number.of.knots.pre.break[1] <= 0)
stop('"prior.for.number.of.knots.pre.break[1]" must be greater than zero')
if(prior.for.number.of.knots.pre.break[2] <= 0 ||prior.for.number.of.knots.pre.break[2] >= 1 )
stop('"prior.for.number.of.knots.pre.break[2]" must be greater than zero but less than one')
ans@priorForNumberOfKnotsPreBreak <- prior.for.number.of.knots.pre.break
if(!is.numeric(prior.for.number.of.knots.post.break) || length(prior.for.number.of.knots.post.break) != 2)
stop('"prior.for.number.of.knots.post.break" must be a numeric of length 2')
if(prior.for.number.of.knots.post.break[1] <= 0)
stop('"prior.for.number.of.knots.post.break[1]" must be greater than zero')
if(prior.for.number.of.knots.post.break[2] <= 0 ||prior.for.number.of.knots.post.break[2] >= 1 )
stop('"prior.for.number.of.knots.post.break[2]" must be greater than zero but less than one')
ans@priorForNumberOfKnotsPostBreak <- prior.for.number.of.knots.post.break
if(sum(ans@exposureYears < ans@rangeForFirstYearInNewRegime[1]) < 4)
stop('The minimum value for "first.year.in.new.regime" is too small. There must be at least 4 years in the pre-break period.')
if(sum(ans@exposureYears >= ans@rangeForFirstYearInNewRegime[2]) < 4)
stop('The maximum value for "first.year.in.new.regime" is too large. There must be at least 4 years in the post-break period.')
if(!is.numeric(prior.for.first.year.in.new.regime) || length(prior.for.first.year.in.new.regime) != 2)
stop('The value supplied for "prior.for.first.year.in.new.regime" must be a numeric vector of length 2.')
if(any(prior.for.first.year.in.new.regime <=0))
stop('"prior.for.first.year.in.new.regime" must be greater than 0')
ans@priorsForFirstYearInNewRegime <- prior.for.first.year.in.new.regime
ans@outputType <- 'BreakAnnualAggLossDevModelOutput'
ans@modelFile <- 'break.model.txt'
return(invisible(ans))
}
##' A method to collect all the needed model input specific to the break model.
##'
##' There are currently two types of \code{AnnualAggLossDevModel}s (break and standard). These models have many data elements in common.
##' This method appends only the elements specific to the break model onto the list created by a call to \code{NextMethod()}.
##'
##' The following elements are appended:
##' \describe{
##' \item{\code{x.0}}{Two values. The lower bound for the number of knots. First is for first spline.}
##' \item{\code{x.r}}{Two values. The upper bound for the number of knots. First is for first spline.}
##' \item{\code{break.row}}{Two integer values given the (inclusive) range of rows in which the first year in the new regime could occur.}
##' \item{\code{break.row.priors}}{The parameters for the beta distribution which serves as the prior for the location of the structural break.}
##' \item{\code{K.trim}}{The maximum number of columns in the post-break triangle.}
##' \item{\code{beta.prior}}{A matrix giving the prior for the location of knots. First column is for the pre-break spline. Second is for the post-break spline.}
##' \item{\code{mu.number.of.knots.prior}}{A matrix giving the prior for the mean of the number of knots. First column is for the pre-break spline. Second is for the post-break spline}
##' \item{\code{number.of.knots.ubound}}{A vector giving the upper bound for the number of knots. First is for the pre-break spline. Second is for the post-break spline}
##' }
##' @name getJagsData,BreakAnnualLossDevModelInput-method
##' @param object An object of type \code{BreakAnnualAggLossDevModelInput} from which to collect the needed model input.
##' @return A named list of the specific model elements. See details for more info.
##' @docType methods
setMethod(
'getJagsData',
signature(object='BreakAnnualAggLossDevModelInput'),
function(object)
{
ans <- callNextMethod()
##return(ans)
log.inc.index <- ans$log.inc.index
##calculate the last column with two non-missing values
x.r <- function()
{
log.inc <- ans$log.inc
for(i in dim(log.inc)[2]:1)
{
tmp <- sum(!is.na(log.inc[,i]))
if(tmp >= 2)
return(i)
}
stop('"log.inc" does not have a column with at least two non missing values')
}
x.r.post.break <- function()
{
log.inc <- ans$log.inc
log.inc <- log.inc[object@exposureYears >= max(object@rangeForFirstYearInNewRegime),]
for(i in dim(log.inc)[2]:1)
{
tmp <- sum(!is.na(log.inc[,i]))
if(tmp >= 2)
return(i)
}
stop('"log.inc" for the post period does not have a column with at least two non missing values')
}
ans$x.0 <- c(1, 1)
ans$x.r <- c(x.r(), x.r.post.break())
ans$break.row <- match(object@rangeForFirstYearInNewRegime, object@exposureYears)
ans$break.row.priors <- object@priorsForFirstYearInNewRegime
ans$K.trim <- ans$K - ans$break.row[1] + 1
ans$beta.prior <- array(NA, c(2,2))
ans$beta.prior[,1] <- c(1,object@priorForKnotPositionsPreBreak)
ans$beta.prior[,2] <- c(1,object@priorForKnotPositionsPostBreak)
ans$mu.number.of.knots.prior <- array(NA, c(2, 2))
ans$mu.number.of.knots.prior[,1] <- object@priorForNumberOfKnotsPreBreak
ans$mu.number.of.knots.prior[2,1] <- (1 - ans$mu.number.of.knots.prior[2,1]) / ans$mu.number.of.knots.prior[2,1]
ans$mu.number.of.knots.prior[,2] <- object@priorForNumberOfKnotsPostBreak
ans$mu.number.of.knots.prior[2,2] <- (1 - ans$mu.number.of.knots.prior[2,2]) / ans$mu.number.of.knots.prior[2,2]
ans$number.of.knots.ubound <- numeric(2)
ans$number.of.knots.ubound[1] <- trunc(ans$K/2) + 1
ans$number.of.knots.ubound[2] <- trunc(ans$K.trim/2) + 1
return(ans)
})
##' A method to collect all the needed model initial values unique to the break model.
##'
##' There are currently two types of \code{AnnualAggLossDevModel}s (break and standard). Code needed to create initial values specific the break model is placed in this method.
##' This method returns a parameterless function which when called first calls the function returned by \code{NextMethod()} and then to the list returned by that function appends the following initial values.
##'
##' \describe{
##' \item{\code{R.}}{The initial values for the spline node. Needed because dspline cannot create initial values.}
##' }
##' @name getJagsInits,BreakAnnualLossDevModelInput-method
##' @param object An object of type \code{BreakAnnualAggLossDevModelInput} from which to collect the needed model initial values.
##' @return A named list of the specific model elements. See details for more info.
##' @docType methods
##' @seealso \code{\link{getJagsInits}}
setMethod(
'getJagsInits',
signature(object='BreakAnnualAggLossDevModelInput'),
function(object)
{
super.f <- callNextMethod()
K <- getTriDim(object)[1]
R. <- array(0, c(K, 6))
R.[,2] <- 1
function()
{
ans <- super.f()
ans$R. <- R.
return(ans)
}
})
Try the BALD 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.