Reference

Implements the CreditMetrics model for portfolio credit risk

AbstractProbabilityVector{T}

Abstract type for probability vectors with entries of type T summing to one.

AbstractRealLineThresholds{T}

Abstract type for a weakly sorted vector of thresholds of type T where values are unconstrained

Abstract type for link functions between probabilities and values on the whole real line

AbstractUnitIntervalThresholds{T}

Abstract type for a weakly sorted vector of thresholds of type T where values are between zero and one.

DeformedRealLineThresholds{T, TBase <: AbstractRealLineThresholds{T}} <: AbstractRealLineThresholds{T}

A weakly sorted vector of values on the real line that has been affinely deformed.

DeformedRealLineThresholds holds the specification of the affine deformation and defers calculation of the deformed thresholds until they are needed.

Normal distribution link specification using StatsFuns implementation of normal cdf and invcdf

ProbabilityVector{T, TVec <: AbstractVector{T}} <: AbstractProbabilityVector{T}

Trusted vector of probabilities of type T that sum to one.

ProbabilityVector(ps::TVec) where {T, TVec<:AbstractVector{T}}

Convert a vector of probabilities to a ProbabilityVector, thereby ensuring that probabilities are non-negative and sum to one (trusted). Regarding the last requirement, this constructor is very unforgiving.

ProbabilityVector(apv::AbstractProbabilityVector{T}) where {T}
ProbabilityVector(auit::AbstractUnitIntervalThresholds{T}) where {T}

Convert to ProbabilityVector other types that are trusted.

RealLineThresholds{T, TVec <: AbstractVector{T}} <: AbstractRealLineThresholds{T}

A weakly sorted vector of values on the real line.

The UnitIntervalThresholds struct is used to check that values are weakly sorted, and using the type system to express that.

RealLineThresholds(rlt::TVec) where {T, TVec <: AbstractVector{T}}

Convert a raw vector of thresholds rlt of type T to a RealLineThresholds

RealLineThresholds(arlt::AbstractRealLineThresholds{T}) where {T}

Convert an AbstractRealLineThresholds arlt to a RealLineThresholds by extracting the raw thresholds

RealLineThresholds(auit::AbstractUnitIntervalThresholds{TUnit}, atl::AbstractThresholdLink{TUnit, TReal}) where {TUnit, TReal}

Convert an AbstractUnitIntervalThresholds auit to a RealLineThresholds using the link atl

Student's distribution link specification using StatsFuns implementation of Student's cdf and invcdf

UnitIntervalThresholds{T, TVec <: AbstractVector{T}} <: AbstractUnitIntervalThresholds{T}

A weakly sorted vector of values on the interval from zero and one.

The UnitIntervalThresholds struct is used to check that values are in interval and weakly sorted, and using the type system to express that.

UnitIntervalThresholds(thr::TVec) where {T, TVec <: AbstractVector{T}}

Convert a vector thr of thresholds of type T to a UnitIntervalThresholds

UnitIntervalThresholds(auit::AbstractUnitIntervalThresholds{T})

Convert an AbstractUnitIntervalThresholds auit to a UnitIntervalThresholds by extracting its thresholds

UnitIntervalThresholds(apv::AbstractProbabilityVector{T}) where {T}

Convert an AbstractProbabilityVector to a UnitIntervalThresholds

UnitIntervalThresholds(arlt::AbstractRealLineThresholds{TReal}, atl::AbstractThresholdLink{TUnit, TReal}) where {TUnit, TReal}

Convert an AbstractRealLineThresholds arlt to a UnitIntervalThresholds using the threshold link atl

cdf(atl::AbstractThresholdLink)

Get the link function from real line to probability for the link specification atl

cdf(ntl::NormalThresholdLink)

Get the link function from real line to probability for ntl which is always the normal cdf

cdf(stl::StudentThresholdLink)

Get the link function from real line to probability for stl which is the Student's cdf with degrees of freedom specified by stl

Make a deformation of probabilities or thresholds

deformation(arlt::AbstractRealLineThresholds, push, norm)
deformation(auit::AbstractUnitIntervalThresholds, atl::AbstractThresholdLink, push, norm)
deformation(apv::AbstractProbabilityVector, atl::AbstractThresholdLink, push, norm)

Deformations are made on the thresholds on the real line. They consist of a push that moves the thresholds and a norm with which the pushed thresholds are divided for proper normalization.

Methods on probabilities and unit interval thresholds are provided for ease of use and also depend on a threshold link atl.

Compute the mean value after a deformation when the values for each outcome is given.

invcdf(atl::AbstractThresholdLink)

Get the link function from probability to real line for the link specification atl

invcdf(ntl::NormalThresholdLink)

Get the link function from probability to real line for ntl which is always the normal inverse cdf

invcdf(stl::StudentThresholdLink)

Get the link function from probability to real line for stl which is the Student's inverse cdf with degrees of freedom specified by stl

Compute the mean value when given the value for each outcome.

Compute the outcome of a particular realization of scenario

Compute the values for one or more outcomes.

probabilities(apv::AbstractProbabilityVector{T})

Get the probabilities of apv in a raw format, typically an AbstractVector{T}.

thresholds(auit::AbstractUnitIntervalThresholds{T}) where {T}

Get the thresholds of auit in a raw format, typically an AbstractVector{T}.

thresholds(arlt::AbstractRealLineThresholds{T}) where {T}

Get the thresholds of arlt in a raw format, typically an AbstractVector{T}.