Rearranged copulas and the Schur order #
For a copula E, E↑(u,v)=∫_0^u (∂₁E(·,v))* is the increasing (SI) rearrangement and
E↓(u,v)=v-E↑(1-u,v) the decreasing one. We prove that E↑ is a copula, CIS, Schur
equivalent to E, and the lower-orthant maximum of {D : D ≤_{∂₁S} E}; E↓ is the
minimum. This gives Lemma 2.7 and Proposition 3.1 of Ansari–Rockel.
The conditional section u ↦ P(V ≤ v | U=u).
Equations
- Verification.condSection C v u = C.conditionalCDF u v
Instances For
The increasing rearranged copula, as a CDF.
Equations
- Verification.upRearrCDF C u v = ∫ (s : ↑unitInterval) in Set.Iic u, Verification.decRearr (Verification.condSection C v) ↑s
Instances For
The increasing rearranged copula C↑.
Equations
- Verification.upRearr C = ProbabilityTheory.Copula.ofClassical (fun (u : Fin 2 → ↑unitInterval) => Verification.upRearrCDF C (u 0) (u 1)) ⋯
Instances For
The decreasing rearranged copula C↓(u,v)=v-C↑(1-u,v).
Equations
Instances For
The conditional distribution of C↑ is the decreasing rearrangement.
Lemma 2.7(ii): C↑ is conditionally increasing (CIS).
Lemma 2.7(iv): C↑ is Schur equivalent to C.
Hardy–Littlewood on arbitrary measurable sets.
The directional Schur order of copulas in its rearrangement form, on conditional CDFs.
Equations
- Verification.CondRearrSchurLE D E = ∀ (v : ↑unitInterval), Verification.RearrSchurLE (Verification.condSection D v) (Verification.condSection E v)
Instances For
Lemma 2.7(i): E↑ is the lower-orthant maximum of {D : D ≤_{∂₁S} E}.
Lemma 2.7(i): E↓ is the lower-orthant minimum of {D : D ≤_{∂₁S} E}.
Proposition 3.1 (i)⇔(ii).
Proposition 3.1 (ii)⇔(iii).