The positive-parameter Clayton construction #
Take independent rate-one exponentials E i and an independent gamma variable
G with shape 1/θ and rate one. The joint law of -E i / G has atomless
marginals; its unique Sklar copula is the gamma-frailty construction of Clayton.
This module supplies the stochastic construction and its Sklar factorization.
Copula.Families.Clayton.CDF identifies the closed-form Archimedean CDF, and
Copula.Families.Clayton.Limits proves its two endpoint limits.
noncomputable def
ProbabilityTheory.Copula.claytonFrailtySource
(d : ℕ)
(θ : ℝ)
(hθ : 0 < θ)
:
MeasureTheory.ProbabilityMeasure ((Fin d → ℝ) × ℝ)
Independent exponential numerators and their common gamma frailty.
Equations
- ProbabilityTheory.Copula.claytonFrailtySource d θ hθ = ⟨(MeasureTheory.Measure.pi fun (x : Fin d) => ProbabilityTheory.expMeasure 1).prod (ProbabilityTheory.gammaMeasure θ⁻¹ 1), ⋯⟩
Instances For
The joint negative exponential/gamma ratios used in the Clayton construction.
Equations
Instances For
theorem
ProbabilityTheory.Copula.atomless_claytonLaw_marginal
(d : ℕ)
(θ : ℝ)
(hθ : 0 < θ)
(i : Fin d)
:
MeasureTheory.NullSingletonClass (marginal (claytonLaw d θ hθ) i)
theorem
ProbabilityTheory.Copula.continuous_claytonLaw_marginal
(d : ℕ)
(θ : ℝ)
(hθ : 0 < θ)
(i : Fin d)
:
Continuous ↑(ProbabilityTheory.cdf (marginal (claytonLaw d θ hθ) i))
The positive-parameter Clayton copula, constructed from gamma frailty.
Equations
Instances For
theorem
ProbabilityTheory.Copula.isSklarCopula_clayton
(d : ℕ)
(θ : ℝ)
(hθ : 0 < θ)
:
IsSklarCopula (claytonLaw d θ hθ) (clayton d θ hθ)