Constructed clamped extremizers and their unique optimality #
Equations
Instances For
The source's normalization parameter, including response thresholds 0 and 1.
Equations
- Papers.Rockel2026XiBlest.extremalQ b hb v = -Verification.quadraticIntercept b ⋯ v / b
Instances For
theorem
Papers.Rockel2026XiBlest.extremal_kernel_measurable
(b : ℝ)
(hb : 0 ≤ b)
:
Measurable fun (p : ↑unitInterval × ↑unitInterval) => Verification.quadraticKernel b hb p.2 p.1
theorem
Papers.Rockel2026XiBlest.extremal_conditionalCDF
(b : ℝ)
(hb : 0 < b)
(v : ↑unitInterval)
:
(fun (u : ↑unitInterval) => (extremalCopula b ⋯).conditionalCDF u v) =ᵐ[MeasureTheory.volume] fun (u : ↑unitInterval) =>
Verification.unitClamp (b * ((1 - ↑u) ^ 2 - extremalQ b hb v))
theorem
Papers.Rockel2026XiBlest.extremal_support
(C : ProbabilityTheory.Copula 2)
(b : ℝ)
(hb : 0 ≤ b)
:
b * blestNu C - C.chatterjeeXi ≤ b * blestNu (extremalCopula b hb) - (extremalCopula b hb).chatterjeeXi
theorem
Papers.Rockel2026XiBlest.extremal_support_eq_iff
(C : ProbabilityTheory.Copula 2)
(b : ℝ)
(hb : 0 ≤ b)
:
b * blestNu C - C.chatterjeeXi = b * blestNu (extremalCopula b hb) - (extremalCopula b hb).chatterjeeXi ↔ C = extremalCopula b hb
theorem
Papers.Rockel2026XiBlest.extremal_maximal_blest
(C : ProbabilityTheory.Copula 2)
(b : ℝ)
(hb : 0 < b)
(hx : C.chatterjeeXi ≤ (extremalCopula b ⋯).chatterjeeXi)
:
theorem
Papers.Rockel2026XiBlest.extremal_maximal_blest_eq_iff
(C : ProbabilityTheory.Copula 2)
(b : ℝ)
(hb : 0 < b)
(hx : C.chatterjeeXi = (extremalCopula b ⋯).chatterjeeXi)
: