We study how to perform tests on samples of pairs of observations and predictions in order to assess whether or not the predictions are prudent. Prudence requires that the mean of the difference of the observation-prediction pairs can be shown to be significantly negative. For safe conclusions,we suggest testing both unweighted (or equally weighted) and weighted means and explicitly taking into account the randomness of individual pairs. The test methods presented are mainly specified as bootstrap and normal approximation algorithms. The tests are general but can be applied in particular in the area of credit risk,both for regulatory and accounting purposes.
Citation: Dirk Tasche. Proving prediction prudence[J]. Data Science in Finance and Economics, 2022, 2(4): 335-355. doi: 10.3934/DSFE.2022017
[1] | Wen Teng, Xiansheng Dai . Nonabelian embedding tensors on 3-Lie algebras and 3-Leibniz-Lie algebras. Electronic Research Archive, 2025, 33(3): 1367-1383. doi: 10.3934/era.2025063 |
[2] | Shanshan Liu, Abdenacer Makhlouf, Lina Song . The full cohomology, abelian extensions and formal deformations of Hom-pre-Lie algebras. Electronic Research Archive, 2022, 30(8): 2748-2773. doi: 10.3934/era.2022141 |
[3] | Qiang Li, Lili Ma . 1-parameter formal deformations and abelian extensions of Lie color triple systems. Electronic Research Archive, 2022, 30(7): 2524-2539. doi: 10.3934/era.2022129 |
[4] | Xueru Wu, Yao Ma, Liangyun Chen . Abelian extensions of Lie triple systems with derivations. Electronic Research Archive, 2022, 30(3): 1087-1103. doi: 10.3934/era.2022058 |
[5] | Youming Chen, Weiguo Lyu, Song Yang . A note on the differential calculus of Hochschild theory for A∞-algebras. Electronic Research Archive, 2022, 30(9): 3211-3237. doi: 10.3934/era.2022163 |
[6] | Kailash C. Misra, Sutida Patlertsin, Suchada Pongprasert, Thitarie Rungratgasame . On derivations of Leibniz algebras. Electronic Research Archive, 2024, 32(7): 4715-4722. doi: 10.3934/era.2024214 |
[7] | Hongliang Chang, Yin Chen, Runxuan Zhang . A generalization on derivations of Lie algebras. Electronic Research Archive, 2021, 29(3): 2457-2473. doi: 10.3934/era.2020124 |
[8] | Margarida Camarinha . A natural 4th-order generalization of the geodesic problem. Electronic Research Archive, 2024, 32(5): 3396-3412. doi: 10.3934/era.2024157 |
[9] | Pengliang Xu, Xiaomin Tang . Graded post-Lie algebra structures and homogeneous Rota-Baxter operators on the Schrödinger-Virasoro algebra. Electronic Research Archive, 2021, 29(4): 2771-2789. doi: 10.3934/era.2021013 |
[10] | Xiuhai Fei, Cuixian Lu, Haifang Zhang . Nonlinear Jordan triple derivable mapping on ∗-type trivial extension algebras. Electronic Research Archive, 2024, 32(3): 1425-1438. doi: 10.3934/era.2024066 |
We study how to perform tests on samples of pairs of observations and predictions in order to assess whether or not the predictions are prudent. Prudence requires that the mean of the difference of the observation-prediction pairs can be shown to be significantly negative. For safe conclusions,we suggest testing both unweighted (or equally weighted) and weighted means and explicitly taking into account the randomness of individual pairs. The test methods presented are mainly specified as bootstrap and normal approximation algorithms. The tests are general but can be applied in particular in the area of credit risk,both for regulatory and accounting purposes.
Derivations play a crucial role in studying deformation formulas [1], differential Galois theory [2] and homotopy algebras [3]. They also are useful of control systems theory [4,5] and gauge algebras [6]. The authors studied the operad of associative algebras with derivation in [7]. Recently, the authors introduced Lie algebras with derivations, and studied their cohomology and deformations, extensions in [8]. Later, Das[9] considered the similar results for Leibniz algebras with derivations. The authors studied cohomology of Leibniz triple systems with derivations in [10].
More and more scholars have begun to pay close attention to the structures of any weight thanks to the result of outstanding work [12], all kinds of Rota-Baxter algebras of any weight [13,14,15,16,17,18] appear successively. In order to study non-abelian extensions of Lie algebras. The notion of crossed homomorphisms of Lie algebras was introduced by Lue [19], which was applied to study the representations of Cartan Lie algebras [20]. For λ∈k, the notion of a differential algebra of weight λ was first introduced by Guo and Keigher [21], which generalizes simultaneously the concept of the classical differential algebra and difference algebra [22]. Applying the same method as for differential Lie algebras of weight λ. Later, the authors defined the cohomology of relative difference Lie algebras, and studied some properties in [23]. Our aim in this paper is to consider Lie algebras with differential operators of weight λ (also known as differential Lie algebras). More precisely, we define a cohomology theory for differential Lie algebras and consider some properties.
The paper is organized as follows. In Section 2, we consider the representations of differential Lie algebras of any weight. In Section 3, we define a cohomology theory for differential Lie algebras of any weight. In Section 4, we study central extensions of differential Lie algebras of any weight. In Section 5, we study formal deformations of differential Lie algebras of any weight. In Section 6, we consider homotopy differential operators on L∞ algebras and 2-differential operators of any weight on Lie 2-algebras. In Section 7, we prove that the category of 2-term L∞ algebras with homotopy differential operator of any weight and the category of Lie 2-algebras with 2-differential operators of any weight are equivalent.
Throughout this paper, k denotes a field of characteristic zero. All the vector spaces, algebras, linear maps and tensor products are taken over k unless otherwise specified.
For λ∈k. A differential operator of weight λ on a Lie algebra g is a linear operator dg:g→g such that
dg([a,b])=[dg(a),b]+[a,dg(b)]+λ[dg(a),dg(b)],∀a,b∈g. | (2.1) |
We denote by Derλ(g) the set of differential operators of weight λ of the Lie algebra g.
Definition 2.1. Denote a Lie algebra g with a differential operator dg∈Derλ(g) by (g,dg) and we call it a differential Lie algebra.
Definition 2.2. Given two differential Lie algebras (g,dg),(h,dh), a homomorphism of differential Lie algebras is a Lie algebra homomorphism φ:g→h such that φ∘dg=dh∘φ. We denote by LieDλ the category of differential Lie algebras and their morphisms.
To simply notations, for all the above notions, we will often suppress the mentioning of the weight λ unless it needs to be specified.
Definition 2.3. (i) A representation over the differential Lie algebra (g,dg) is a pair (V,dV), where dV∈Endk(V), and (V,ρ) is a representation over the Lie algebra g, such that ∀x∈g,v∈V, the following identity holds:
dV(ρ(x)v)=ρ(dg(x))v+ρ(x)dV(v)+λρ(dg(x))dV(v). |
(ii) Given two representations (U,ρU,dU),(V,ρV,dV) over (g,dg), a linear map f:U→V is called a homomorphism of representations, if f∘dU=dV∘f and
f∘ρU(a)=ρV(a)∘f,∀a∈g. |
One denotes by (g,dg)-Rep the category of representations over the differential Lie algebra (g,dg).
Example 2.4. Any differential Lie algebra (g,dg) is a representation over itself with
ρ:g→Endk(g),a↦(b↦[a,b]). |
It is called the adjoint representation over the differential Lie algebras (g,dg).
Example 2.5. Let (V,ρ) be a representation of a Lie algebra g. Then the pair (V, IdV) is a representation of the differential Lie algebra (g,Idg) of weight −1.
Example 2.6. Let (g,dg) be a differential Lie algebra of weight λ and (V,dV) be a representation of it. Then for κ≠0∈k, the pair (V,κdV) is a representation of the differential Lie algebra (g,κdg) of any weight 1κλ.
The following result is easily to check and we omit it.
Proposition 2.7. Let (V,dV) be a representation of the differential Lie algebra (g,dg) of weight λ. Then (g⊕V,dg⊕dV) is a differential Lie algebra, where
[a+u,b+v]=[a,b]+ρ(a)v−ρ(b)u,∀a,b∈g,u,v∈V. |
Recall that the cochain complex of Lie algebra g with coefficients in representation V is the cochain complex
(C∗Lie(g,V)=⊕∞n=0CnLie(g,V),∂∗Lie), |
and the coboundary operator
∂nLie:CnLie(g,V)⟶Cn+1Lie(g,V),n⩾0 |
is given by
∂nLief(x1,…,xn+1)=n+1∑i=1(−1)i+nρ(xi)f(x1,…,ˆxi,…,xn+1)+n∑1≤i<j≤n+1(−1)i+j+n+1f([xi,xj],x1,…,ˆxi,…,ˆxj,…,xn+1), |
for all f∈CnLie(g,V),x1,…,xn+1∈g. The corresponding cohomology is denoted by H∗Lie(g,V). When V is the adjoint representation, we write HnLie(g)=HnLie(g,V),n≥0.
In the following, we will define the cohomology of the differential Lie algebra (g,dg) of weight λ with coefficients in the representation (V,dV).
Define
CnLieDλ(g,V):={CnLie(g,V)⊕Cn−1Lie(g,V),n≥2,C1Lie(g,V)=Hom(g,V),n=1,C0Lie(g,V)=V,n=0. | (3.1) |
and define a linear map δ:CnLie(g,V)→CnLie(g,V) (n≥1) by
δfn(x1,…,xn):=n∑k=1λk−1∑1≤i1<⋯<ik≤nfn(x1,…,dg(xi1),…,dg(xik),…,xn)−dVfn(x1,…,xn), |
for any fn∈CnLieDλ(g,V) and
δv=−dV(v),∀v∈C0LieDλ(g,V)=V. |
Lemma 3.1. We have ∂Lie∘δ=δ∘∂Lie.
Theorem 3.2. The pair (C∗LieDλ(g,V),∂LieDλ) is a cochain complex. So ∂2LieDλ=0.
Proof. For any v∈C0LieDλ(g,V), we have
∂2LieDλv=∂LieDλ(∂LieDλv,δv)=(∂2Liev,∂Lieδv−δ∂Liev)=0. |
Given any f∈CnLie(g,V),g∈Cn−1Lie(g,V) with n≥1, we have
∂2LieDλ(f,g)=∂LieDλ(∂Lief,∂Lieg+(−1)nδf)=(∂2Lief,∂Lie(∂Lieg+(−1)nδf)+(−1)n+1δ∂Lief)=0. |
Hence, the proof is finished.
Definition 3.3. The cohomology of the cochain complex (C∗LieDλ(g,V),∂LieDλ), denoted by H∗LieDλ(g,V), is called the cohomology of the differential Lie algebra (g,dg) of weight λ.
In this section, we show that abelian extensions of differential Lie algebras are classified by the second cohomology.
Definition 4.1. An abelian extension of differential Lie algebras is a short exact sequence of homomorphisms of differential Lie algebras
![]() |
such that [u,v]V=0 for all u,v∈V.
(ˆg,dˆg) is called an abelian extension of (g,dg) by (V,dV).
Definition 4.2. Let (ˆg1,dˆg1) and (ˆg2,dˆg2) be two abelian extensions of (g,dg) by (V,dV). They are said to be isomorphic if there exists ζ:(ˆg1,dˆg1)↔(ˆg2,dˆg2) is an isomorphism of differential Lie algebras such that:
![]() |
A section of an abelian extension (ˆg,dˆg) of (g,dg) by (V,dV) is a linear map s:g→ˆg satisfying p∘s=Idg.
Given a section s:g→ˆg, define ρ:g→Endk(V) by
ρ(x)v:=ρ(s(x))v,∀x∈g,v∈V. |
Proposition 4.3. With the above notations, (V,dV) is a representation over the differential Lie algebra (g,dg).
Proof. Firstly, we prove that ρ is a Lie algebra homomorphism, in fact, for any x,y∈g,v∈V, we have
ρ([x,y])(v)=ρ(s([x,y]))v=ρ([s(x),s(y)])v=ρ(x)ρ(y)(v)−ρ(y)ρ(x)(v). |
Moreover, we obtain
dV(ρ(x)v)=dV(ρ(s(x))v)=dˆg(ρ(s(x))v)=ρ(dˆg)(s(x))v+ρ(s(x))dˆg(v)+λρ(dˆg(s(x)))dˆg(v)=ρ(s(dg(x)))v+ρ(s(x))dV(v)+λρ(s(dg(x)))dV(v)=ρ(dg(x))v+ρ(x)dV(v)+λρ(dg(x))dV(v). |
Hence, (V,dV) is a representation over (g,dg).
We further consider linear maps ψ:g⊗g→V and χ:g→V by
ψ(x,y)=[s(x),s(y)]−s([x,y]),∀x,y∈g,χ(x)=dˆg(s(x))−s(dg(x)),∀x∈g. |
The differential Lie algebra structure on g⊕V with a multiplication [⋅,⋅]ψ and the differential operator dχ defined by
[x+u,y+v]ψ=[x,y]+ρ(x)v−ρ(y)u+ψ(x,y),∀x,y∈g,u,v∈V, | (4.1) |
dχ(x+v)=dg(x)+χ(x)+dV(v),∀x∈g,v∈V. | (4.2) |
Proposition 4.4. The triple (g⊕V,[⋅,⋅]ψ,dχ) is a differential Lie algebra if and only if (ψ,χ) is a 2-cocycle.
Proof. For any x,y,z∈g, By (4.1), we have
ψ(x,[y,z])+ψ(x,ψ(y,z))+ψ(y,[z,x])+ψ(y,ψ(z,x))+ψ(z,[x,y])+ψ(z,ψ(x,y))=0. | (4.3) |
Since dχ satisfies Eq (2.1), we deduce that
χ([x,y])−ρ(x)χ(y)−λρ(dg(x))χ(y)+ρ(y)χ(x)+λρ(dg(y))χ(x)+dV(ψ(x,y))−ψ(dg(x),y)−ψ(x,dg(y))−λψ(dg(x),dg(y))=0. | (4.4) |
Therefore, (ψ,χ) is a 2-cocycle.
Conversely, if (ψ,χ) satisfies Eqs (4.3) and (4.4), direct verification that (g⊕V,[⋅,⋅]ψ,dχ) is a differential Lie algebra.
In the following, we will classify abelian extensions of differential Lie algebras.
Theorem 4.5. Let V be a vector space and dV∈Endk(V). Then abelian extensions of a differential Lie algebra (g,dg) by (V,dV) are classified by H2LieDλ(g,V) of (g,dg).
Proof. Let (ˆg,dˆg) be an abelian extension of (g,dg) by (V,dV). We choose a section s:g→ˆg to obtain a 2-cocycle (ψ,χ) and let s1 and s2 be two distinct sections providing 2-cocycles (ψ1,χ1) and (ψ2,χ2) respectively. Define ϕ:g→V by ϕ(x)=s1(x)−s2(x), we have
ψ1(x,y)=[s1(x),s1(y)]−s1([x,y])=[s2(x)+ϕ(x),s2(y)+ϕ(y)]−(s2([x,y])+ϕ([x,y]))=([s2(x),s2(y)]−s2([x,y]))+[s2(x),ϕ(y)]+[ϕ(x),s2(y)]−ϕ([x,y])=([s2(x),s2(y)]−s2([x,y]))+[x,ϕ(y)]+[ϕ(x),y]−ϕ([x,y]=ψ2(x,y)+∂ϕ(x,y) |
and
χ1(x)=dˆg(s1(x))−s1(dg(x))=dˆg(s2(x)+ϕ(x))−(s2(dg(x))+ϕ(dg(x)))=(dˆg(s2(x))−s2(dg(x)))+dV(ϕ(x))−ϕ(dg(x))=χ2(x)+dV(ϕ(x))−ϕ(dg(x))=χ2(x)−δϕ(x). |
That is, (ψ1,χ1)=(ψ2,χ2)+∂LieDλ(ϕ). Thus (ψ1,χ1) and (ψ2,χ2) are in the same cohomological class in H2LieDλ(g,V).
Next, we prove that isomorphic abelian extensions give rise to the same element in H2LieDλ(g,V). Assume that (ˆg1,dˆg1) and (ˆg2,dˆg2) are two isomorphic abelian extensions of (g,dg) by (V,dV) with the associated homomorphism ζ:(ˆg1,dˆg1)→(ˆg2,dˆg2). Let s1 be a section of (ˆg1,dˆg1). As p2∘ζ=p1, we have
p2∘(ζ∘s1)=p1∘s1=Idg. |
Therefore, ζ∘s1 is a section of (ˆg2,dˆg2). Denote s2:=ζ∘s1. Since ζ is a homomorphism of differential Lie algebras such that ζ|V=IdV, we have
ψ2(x,y)=[s2(x),s2(y)]−s2([x,y])=[ζ(s1(x)),ζ(s1(y))]−ζ(s1([x,y]))=ζ([s1(x),s1(y)]−s1([x,y]))=ζ(ψ1(x,y))=ψ1(x,y) |
and
χ2(x)=dˆg2(s2(x))−s2(dg(x))=dˆg2(ζ(s1(x)))−ζ(s1(dg(x)))=ζ(dˆg1(s1(x))−s1(dg(x)))=ζ(χ1(x))=χ1(x). |
Therefore, the result can be obtained.
Conversely, given two 2-cocycles (ψ1,χ1) and (ψ2,χ2), we can construct two abelian extensions (g⊕V,[⋅,⋅]ψ1,dχ1) and (g⊕V,[⋅,⋅]ψ2,dχ2) via Eqs (4.1) and (4.2), and then there exists a linear map ϕ:g→V such that
(ψ1,χ1)=(ψ2,χ2)+∂LieDλ(ϕ). |
Define ζ:g⊕V→g⊕V by
ζ(x,v):=(x,ϕ(x)+v). |
Then ζ is an isomorphism of these two abelian extensions.
In this section, we show that if H2LieDλ(g,g)=0, then the differential Lie algebra (g,dg) is rigid.
Let (g,dg) be a differential Lie algebra. Denote by μg the multiplication of g. Consider the 1-parameterized family
μt=∞∑i=0μiti,μi∈C2Lie(g,g),dt=∞∑i=0diti,di∈C1Lie(g,g). |
Definition 5.1. A 1-parameter formal deformation of a differential Lie algebra (g,dg) is a pair (μt,dt) which endows the k[[t]]-module (g[[t]],μt,dt) with the differential Lie algebra over k[[t]] such that (μ0,d0)=(μg,dg).
Given any differential Lie algebra (g,dg), interpret μg and dg as the formal power series μt and dt with μi=δi,0μg and di=δi,0dg respectively for all i≥0. Then (g[[t]],μg,dg) is a 1-parameter formal deformation of (g,dg).
The pair (μt,dt) generates a 1-parameter formal deformation of the differential Lie algebra (g,dg) if and only if the following identities hold:
0=μt(x,μt(y,z))+μt(y,μt(z,x))+μt(z,μt(x,y)), | (5.1) |
dt(μt(x,y))=μt(dt(x),y)+μt(x,dt(y))+λμt(dt(x),dt(y)),∀x,y,z∈g. | (5.2) |
Expanding these identities and collecting coefficients of tn, we see that Eqs (5.1) and (5.2) are equivalent to the systems of identities:
0=n∑i=0μi(x,μn−i(y,z))+μi(y,μn−i(z,x))+μi(z,μn−i(x,y)), | (5.3) |
∑k,l≥0k+l=ndlμk(x,y)=∑k,l≥0k+l=n(μk(dl(x),y)+μk(x,dl(y)))+λ∑k,l,m≥0k+l+m=nμk(dl(x),dm(y)). | (5.4) |
Remark 5.2. For n=0, Eq (5.3) is equal to the Jabobi identity of μg, and Eq (5.4) is equal to the fact that dg is a differential operator of weight λ.
Proposition 5.3. Let (g[[t]],μt,dt) be a 1-parameter formal deformation of a differential Lie algebra (g,dg). Then (μ1,d1) is a 2-cocycle of the differential Lie algebra (g,dg) with the coefficient in the adjoint representation (g,dg).
Proof. For n=1, Eq (5.3) is equal to ∂Lieμ1=0, and Eq (5.4) is equal to
∂Lied1+δμ1=0. |
Thus for n=1, Eqs (5.3) and (5.4) imply that (μ1,d1) is a 2-cocycle.
If μt=μg in the above 1-parameter formal deformation of the differential Lie algebra (g,dg), we obtain a 1-parameter formal deformation of the differential operator dg. Consequently, we have
Corollary 5.4. Let dt be a 1-parameter formal deformation of the differential operator dg. Then d1 is a 1-cocycle of the differential operator dg with coefficients in the adjoint representation (g,dg).
Proof. In the special case when n=1, Eq (5.4) is equal to ∂Lied1=0, which implies that d1 is a 1-cocycle of the differential operator dg with coefficients in the adjoint representation (g,dg).
Definition 5.5. The 2-cocycle (μ1,d1) is called the infinitesimal of the 1-parameter formal deformation (g[[t]],μt,dt) of (g,dg).
Definition 5.6. Two 1-parameter formal deformations (g[[t]],μt,dt) and (g[[t]],ˉμt,ˉdt) of (g,dg) are said to be equivalent if there exists a formal isomorphism from (g[[t]],ˉμt,ˉdt) to (g[[t]],μt,dt) is a power series Φt=∑i≥0ϕiti:g[[t]]→g[[t]], where ϕi:g→g are linear maps with ϕ0=Idg, such that
Φt∘ˉμt=μt∘(Φt×Φt), | (5.5) |
Φt∘ˉdt=dt∘Φt. | (5.6) |
Theorem 5.7. The infinitesimals of two equivalent 1-parameter formal deformations of (g,dg) are in the same cohomology class H2LieDλ(g,g).
Proof. Let Φt:(g[[t]],ˉμt,ˉdt)→(g[[t]],μt,dt) be a formal isomorphism. For all x,y∈g, we have
Φt∘ˉμt(x,y)=μt∘(Φt×Φt)(x,y),Φt∘ˉdt(x)=dt∘Φt(x). |
Furthermore, we obtain
ˉμ1(x,y)=μ1(x,y)+[ϕ1(x),y]+[x,ϕ1(y)]−ϕ1([x,y]),ˉd1(x)=d1(x)+dg(ϕ1(x))−ϕ1(dg(x)). |
Thus, we have
(ˉμ1,ˉd1)=(μ1,d1)+∂LieDλ(ϕ1), |
which implies that [(ˉμ1,ˉd1)]=[(μ1,d1)] in H2LieDλ(g,g).
Definition 5.8. A 1-parameter formal deformation (g[[t]],μt,dt) of (g,dg) is said to be trivial if it is equal to the deformation (g[[t]],μg,dg), that is, there exists Φt=∑i≥0ϕiti:g[[t]]→g[[t]], where ϕi:g→g are linear maps with ϕ0=Idg, such that
Φt∘μt=μg∘(Φt×Φt), | (5.7) |
Φt∘dt=dg∘Φt. | (5.8) |
Definition 5.9. A differential Lie algebra (g,dg) is said to be rigid if every 1-parameter formal deformation is trivial.
Theorem 5.10. Regarding (g,dg) as the adjoint representation over itself, if H2LieDλ(g,g)=0, the differential Lie algebra (g,dg) is rigid.
Proof. Let (g[[t]],μt,dt) be a 1-parameter formal deformation of (g,dg). By Proposition 5.3, (μ1,d1) is a 2-cocycle. By H2LieDλ(g,g)=0, there exists a 1-cochain ϕ1∈C1Lie(g,g) such that
(μ1,d1)=−∂LieDλ(ϕ1). | (5.9) |
Then setting Φt=Idg+ϕ1t, we have a deformation (g[[t]],ˉμt,ˉdt), where
ˉμt(x,y)=(Φ−1t∘μt∘(Φt×Φt))(x,y),ˉdt(x)=(Φ−1t∘dt∘Φt)(x). |
Thus, (g[[t]],ˉμt,ˉdt) is equivalent to (g[[t]],μt,dt). Furthermore, we have
ˉμt(x,y)=(Idg−ϕ1t+ϕ21t2+⋯+(−1)iϕi1ti+⋯)(μt(x+ϕ1(x)t,y+ϕ1(y)t)),ˉdt(x)=(Idg−ϕ1t+ϕ21t2+⋯+(−1)iϕi1ti+⋯)(dt(x+ϕ1(x)t)). |
Therefore,
ˉμt(x,y)=[x,y]+(μ1(x,y)+[x,ϕ1(y)]+[ϕ1(x),y]−ϕ1([x,y]))t+ˉμ2(x,y)t2+⋯,ˉdt(x)=dg(x)+(dg(ϕ1(x))+d1(x)−ϕ1(dg(x)))t+ˉd2(x)t2+⋯. |
By Eq (5.9), we have
ˉμt(x,y)=[x,y]+ˉμ2(x,y)t2+⋯,ˉdt(x)=dg(x)+ˉd2(x)t2+⋯. |
Then by repeating the argument, we can show that (g[[t]],μt,dt) is equivalent to (g[[t]],μg,dg). Thus, (g,dg) is rigid.
In this section, we pay our attention to the homotopy differential operator of any weight on 2-term L∞-algebras introduced by [24].
Definition 6.1. A 2-term L∞-algebra consists of
● a complex of vector spaces L1d⟶L0,
● bilinear maps l2:Li⊗Lj→Li+j, where i+j≤1,
● a skew-symmetric trilinear map l3:L0⊗L0⊗L0→L1, satisfying:
(a)l2(a,b)=−l2(b,a),l2(a,u)=−l2(u,a),(b)dl2(a,u)=l2(a,du),l2(du,v)=l2(u,dv),(c)dl3(a,b,c)=l2(l2(a,b),c)−l2(l2(a,c),b)−l2(a,l2(b,c)),(d)l3(a,b,du)=l2(l2(a,b),u)−l2(a,l2(b,u))−l2(l2(a,u),b),(e)l2(xa,l3(b,c,w))+l2(l3(a,c,w),b)−l2(l3(a,b,w),c)+l2(l3(a,b,c),w)=l3(l2(a,b),c,w)−l3(l2(a,c),b,w)+l3(l2(a,w),b,c)+l3(a,l2(b,c),w)+l3(a,l2(b,w),c)+l3(a,b,l2(c,w)). |
for any a,b,c,w∈L0 and u,v∈L1.
One denotes a 2-term L∞-algebra as above by (L1d⟶L0,l2,l3). A 2-term L∞-algebra is called skeletal if d=0.
Definition 6.2. Let L=(L1d⟶L0,l2,l3) and L′=(L′1d′⟶L′0,l′2,l′3) be two 2-term L∞-algebras. A morphism f:L→L′ consists of
● a chain map f:L→L′ (which consists of linear maps f0:L0→L′0 and f1:L1→L′1 with f0∘d=d′∘f1),
● a bilinear map f2:L0⊗L0→L′1 satisfying
(a)d(f2(a,b))=f0(l2(a,b))−l′2(f0(a),f0(b)),(b)f2(a,du)=f1(l2(a,u))−l′2(f0(a),f1(u)),(c)f1(l3(a,b,c))+l′2(f0(a,b),f0(c))−l′2(f2(a,c),f0(b))−l′2(f0(a),f2(b,c))+f2(l2(a,b),c)−f2(l2(a,c),b)−f2(a,l2(b,c))−l′3(f0(a),f0(b),f0(c))=0, |
for any a,b,c∈L0 and u∈L1.
If f=(f0,f1,f2):L→L′ and g=(g0,g1,g2):L′→L″ are two morphism of 2-term L∞-algebras, their composition g∘f:L→L″ is defined by (g∘f)0=g0∘f0,(g∘f)1=g1∘f1 and
(g∘f)2(a,b)=g2(f0(a),f0(b))+g1(f2(a,b)),∀a,b∈L0. |
For any 2-term L∞-algebra L, the identity morphism IdL:L→L is given by the identity chain map L→L together with (IdL)2=0.
The collection of 2-term L∞-algebras and morphisms between them form a category. We denote this category by 2Lie∞.
Definition 6.3. Let L=(L1d⟶L0,l2,l3) be a 2-term L∞-algebra. A homotopy differential operator of weight λ on it consists of a chain map of the underlying chain complex (i.e., linear maps θ0:L0→L0 and θ1:L1→L1 with θ0∘d=d∘θ1) and a bilinear map θ2:L0⊗L0→L1 such that for any a,b,c∈L0 and u∈L1, the following identities are hold
(a)d(θ2(a,b))=θ0(l2(a,b))−l2(θ0(a),b)−l2(a,θ0(b))−λl2(θ0(a),θ0(b)),(b)θ2(a,du)=θ1(l2(a,u))−l2(θ0(a),u)−l2(a,θ1(u))−λl2(θ0(a),θ1(u)),(c)l3(θ0(a),b,c)+l3(a,θ0(b),c)+l3(a,b,θ0(c))−θ1(l3(a,b,c))=l2(θ2(a,b),c)−l2(θ2(a,c),b)−l2(a,θ2(b,c))+θ2(l2(a,b),c)−θ2(l2(a,c),b)−θ2(a,l2(b,c)). |
A 2-term L∞-algebra with a homotopy differential operator of weight λ as above denoted by the pair ((L1d⟶L0,l2,l3),(θ0,θ1,θ2)). A 2-term L∞-algebra with a homotopy differential operator of weight λ is said to be skeletal if the underlying 2-term L∞-algebra is skeletal, i.e., d=0.
Definition 6.4. Let ((L1d⟶L0,l2,l3),(θ0,θ1,θ2)) and ((L′1d′⟶L′0,l′2,l′3),(θ′0,θ′1,θ′2)) be two 2-term L∞-algebras with homotopy differential operators of weight λ. A morphism between them consists of a morphism (f0,f1,f2) between the underlying 2-term L∞-algebras and a linear map Ψ:L0→L′1 satisfying
(1)Ψ∘ϕ0=ϕ′1∘Ψ,(2)f0(θ0(a))−θ′0(f0(a))=d′(Ψ(a)),(3)f1(θ1(u))−θ′1(f1(u))=Ψ(da),(4)f1(θ2(a,b))−θ′2(f0(a),f0(b))=θ′1(f2(a,b))−f2(θ0(a),b)−f2(a,θ0(b))+Ψ(l2(a,b))−l′2(Ψ(a),f0(b))−l′2(f0(a),Ψ(b)). |
We denote the category of 2-term L∞-algebras with homotopy differential operators of weight λ and morphisms between them by 2LieDλ∞.
Theorem 6.5. There is a one-to-one correspondence between skeletal 2-term L∞-algebras with homotopy differential operators with weight λ and tuples ((g,dg),(V,dV),(θ,¯θ)), where (g,dg) is a differential Lie algebra of weight λ, (V,dV) is a representation and (θ,¯θ) is a 3-cocycle of the differential Lie algebra of weight λ with coefficients in the representation.
Proof. Let (L10⟶L0,l2,l3,(θ0,θ1,θ2)) be a skeletal 2-term L∞-algebra with a homotopy differential operator of weight λ. Then θ0 is a differential operator of weight λ for the Lie algebra (L0,l2). We have that (L1,θ1) is a representation of the differential Lie algebra (L0,θ0) of weight λ from Definition 6.3. According to the condition (c) in Definition 6.3, we have ∂LieDλ(θ2)+δ(l3)=0. Therefore (l3,−θ2) is a 3-cocycle.
Conversely, define L0=L,L1=V and θ0=dg,θ1=dV,θ2=−¯θ. We define multiplications l2:Li⊗Lj→Li+j and l3:L0⊗L0⊗L0→L1 by
l2(a,b)=[a,b],l2(a,u)=[a,u],l2(u,a)=[u,a],l3=0, |
for a,b,c∈L0=L and u∈L1=V. Then it is easy to verify that ((L10⟶L0,l2,l3),(θ0,θ1,θ2)) is a skeletal 2-term L∞-algebra with a homotopy differential operator of weight λ. Hence, the proof is finished.
A 2-term L∞-algebra with a homotopy differential operator of weight λ is said to be strict if the underlying 2-term L∞-algebra is strict, i.e., θ2=0. Next we introduce crossed modules of differential Lie algebras of weight λ and show that strict 2-term L∞-algebra with a homotopy differential operator of weight λ are in one-to-one correspondence with crossed module of differential Lie algebras of weight λ.
Definition 6.6. A crossed module of differential Lie algebras of weight λ consist of ((g,dg),(h,dh),dt,Λ) where (g,dg) and (h,dh) are differential Lie algebras of weight λ, dt:g→h is a differential Lie algebra morphism and
Λ:h→gl(g),a↦Λa, |
such that for u,v∈g,a,b∈h,
(a)dt(Λa(u))=[a,dt(u)]h,(b)Λdt(u)(v)=[u,v]g,(c)Λ[a,b]h=ΛaΛb−ΛbΛa,(d)dg(Λa(u))=Λdh(a)(u)+Λa(dg(u))+λΛdh(a)(dg(u)). |
Theorem 6.7. There is a one-to-one correspondence between strict 2-term L∞-algebras with homotopy differential operators of weight λ and crossed module of differential Lie algebras of weight λ.
Proof. Let (L1d⟶L0,l2,l3=0,(θ0,θ1,θ2)) be a strict 2-term L∞-algebra with a homotopy differential operator of weight λ. Then θ0 is a differential operator of weight λ for the Lie algebra (L0,l2) and θ1 is a differential operator of weight λ for the Lie algebra (L1,l2) from Definition 6.3. Thus (L0,θ0) and (L1,θ1) are both differential Lie algebras of weight λ. Since θ0∘d=d∘θ1, the map dt=d:L1→L0 is a morphism of differential Lie algebras of weight λ. Finally, the condition (b) of Definition 6.3 is equal to the condition (d) of Definition 6.6. Hence, the results are obtained.
In this section, we study categorified differential operators of any weight (also called 2-differential operator) on Lie 2-algebras.
Definition 7.1. A Lie 2-algebra is a 2-vector space L equipped with
● a bilinear functor [⋅,⋅]:L⊗L→L,
● a trilinear natural isomorphism, called the Jacobiator
Ja,b,c:[[a,b],c]→[[a,c],b]+[a,[b,c]], |
satisfying
![]() |
where Θ,R,P and Q are given by
Θ=J[a,w],b,c+Ja,[b,w],c+Ja,b,[c,w]R=[[[a,w],b],c]+[[a,[b,w]],c]+[[a,b],[c,w]],P=[[[a,c],w],b]+[[a,c],[b,w]]+[[a,w],[b,c]]+[a,[[b,c],w]],Q=[[[a,w],c],b]+[[a,[c,w]],b]+[[a,c],[b,w]]+[[a,w],[b,c]]+[a,[[b,w],c]]+[a,[b,[c,w]]]. |
Definition 7.2. Let (L,[⋅,⋅],J) and (L′,[⋅,⋅]′,J′) be two Lie 2-algebras. A Lie 2-algebra morphism consists of
● a linear functor (F0,F1) from the underlying 2-vector space L to L′;
● a bilinear natural transformation
F2(a,b):[F0(a),F0(b)]′→F0([a,b]) |
satisfying
![]() |
Let L,L′ and L″ be three Lie 2-algebras and F:L→L′,G:L′→L″ be Lie 2-algebra morphisms. Their composition G∘F:L→L″ is a Lie 2-algebra morphism whose components are given by (G∘F)0=G0∘F0,(G∘F)1=G1∘F1 and (G∘F)2 is given by
![]() |
For any Lie 2-algebra L, the identity morphism IdL:L→L is given by the identity functor as its linear functor together with the identity natural transformation as (IdL)2.
Lie 2-algebras and Lie 2-algebra morphisms form a category. We denote this category by Lie2.
In the next, we define 2-differential operators of weight λ on Lie 2-algebras. They are categorification of differential operators on Lie algebras.
Definition 7.3. Let (L,[⋅,⋅],J) be a Lie 2-algebra. A 2-differential operator of weight λ on it consists of a linear map functor D:L→L and a natural isomorphism
Da,b:D[a,b]→[Da,b]+[a,Db]+λ[Da,Db],∀a,b∈L |
satisfying
![]() |
where
P=[D[a,c],b]+[[a,c],D(b)]+[D(a),[b,c]]+[a,D[b,c]]+λ[D[a,c],Db]+λ[Da,D[b,c]]Q=[[Da,c],b]+[[a,Dc],b]+λ[[Da,Dc],b]+[[a,c],D(b)]+[D(a),[b,c]]+[a,[Db,c]]+[a,[b,Dc]]+λ[a,[Db,Dc]]+λ[D[a,c],Db]+λ[Da,D[b,c]]. |
Definition 7.4. Let (L,[⋅,⋅],J,D,D) and (L′,[⋅,⋅]′,J′,D′,D′) be two Lie 2-algebras with 2-differential operators of weight λ. A morphism between them consists of a Lie 2-algebras mophism (F=(F0,F1),F2) and a natural isomorphism
Θa:D′(F0(a))→F0(D(a)),∀a∈L0 |
satisfying
![]() |
We denote the category of Lie 2-algebras with 2-differential operators of weight λ and morphisms between them by LieD2λ.
In the following, we will give our main result of this section.
Theorem 7.5. The categories 2LieDλ∞ and LieD2λ are equivalent.
Proof. First we construct a functor T:2LieDλ∞→LieD2λ as follows. Given a 2-term L∞-algebra with a homotopy differential operator of weight λ ((L1d⟶L0,l2,l3),(θ0,θ1,θ2)), we obtain the 2-vector space C=(L0⊕L1⇉L0). Define a bilinear functor [⋅,⋅]:C⊗C→C by
[(a,u),(b,v)]=(l2(a,b),l2(a,v)+l2(u,b)+l2(du,v)), |
for (a,u),(b,v)∈C1=L0⊕L1. Define
Ja,b,c=([[a,b],c],l3(a,b,c)). |
According to the identities (a)–(e), we can check that (C,[⋅,⋅],J) is a Lie 2-algebra. Moreover, we define a 2-differential operator of weight λ (D,D′) by
D(a,u):=(θ0(a),θ1(u)),Da,b:=([a,b],θ2(a,b)). |
Given any 2-term L∞-algebra with a homotopy differential operator of weight λ morphism (f0,f1,f2,Ψ) from L to L′, for any F0=f0,F1=f1 and
F2(a,b)=([f0(a),f0(b)]′,f2(a,b)),Θ=Ψ. |
Direct verification that F is a morphism from C to C′. Furthermore, we can check that T preserve the identity morphisms and composition of morphisms. Hence, T is a functor from 2LieDλ∞ to LieD2λ.
Conversely. Given a Lie 2-algebra C=(C1⊕C0,J,D,D) with a 2-differential operator of weight λ, we have the 2-term chain complex
L1=kersd=t|kers⟶C0=L0. |
Define l2:Li⊗Lj→Li+j by
l2(a,b)=[a,b],l2(a,u)=[a,u],l2(u,a)=[u,a]. |
The map l3:L0⊗L0⊗L0→L1 is defined by
l3(a,b,c)=pr(Ja,b,c),∀a,b,c∈L0, |
where pr denote the projection on ker(s). Moreover, we define a homotopy differential operator by
θ0(a):=D(i(u)),θ1(u):=D|ker(s)(u),θ2(a,b):=pr(Da,b). |
For any Lie 2-algebra morphism (F0,F1,F2,Θ):C→C′, then f0=F0, f1=F1|L1=kers with a 2-differential operator of weight λ and define f2 by
f2(a,b)=prF2(a,b),Ψ=Θ. |
Moreover, S preserve the identity morphisms and composition of morphisms. Therefore, S is a functor from LieD2λ to 2LieDλ∞.
Finally, it is easy to prove that T∘S≅1LieD2λ, and the composite S∘T≅12LieDλ∞ and we omit them.
The paper is supported by the NSF of China (No. 12271292).
The authors declare no conflict of interest in this paper.
[1] | BCBS. International Convergence of Capital Measurement and Capital Standards. A Revised Framework, Comprehensive Version, 2006. |
[2] | Bellini T(2019) IFRS 9 and CECL Credit Risk Modelling and Validation: A Practical Guide with Examples Worked in R and SAS. Academic Press. |
[3] | Blümke O(2019) Out-of-Time Validation of Default Probabilities within the Basel Accord: A comparative study. SSRN Electron J. http://dx.doi.org/10.2139/ssrn.2945931 |
[4] | Casella G, Berger RL (2002) Statistical Inference. Duxbury Press, second edition. |
[5] | Davison AC, Hinkley DV (1997) Bootstrap Methods and their Application. Cambridge University Press. |
[6] | ECB. Instructions for reporting the validation results of internal models: IRB Pillar I models for credit risk. European Central Bank – Banking Supervision, 2019. |
[7] |
Fischer M, Pfeuffer M (2014) A statistical repertoire for quantitative loss given default validation: overview, illustration, pitfalls and extensions. J Risk Model Validation 8: 3–29. http://dx.doi.org/10.21314/JRMV.2014.115 doi: 10.21314/JRMV.2014.115
![]() |
[8] |
Gürtler M, Hibbeln MT, Usselmann P (2018) Exposure at default modeling – A theoretical and empirical assessment of estimation approaches and parameter choice. J Bank Financ 91: 176–188. http://dx.doi.org/10.1016/j.jbankfin.2017.03.004 doi: 10.1016/j.jbankfin.2017.03.004
![]() |
[9] | Hand DJ (1997) Construction and Assessment of Classification Rules. John Wiley & Sons, Chichester. |
[10] |
Kazianka H (2016) Objective Bayesian estimation of the probability of default. J R Stat Soc Ser C 65: 1–27, 2016. http://dx.doi.org/10.1111/rssc.12107 doi: 10.1111/rssc.12107
![]() |
[11] |
Li D, Bhariok R, Keenan S, et al. (2009) Validation techniques and performance metrics for loss given default models. J Risk Model Validation 3: 3–26. http://dx.doi.org/10.21314/JRMV.2009.045 doi: 10.21314/JRMV.2009.045
![]() |
[12] |
Loterman G, Debruyne M, Vanden BK, et al. (2014) A proposed framework for backtesting loss given default models. J Risk Model Validation 8: 69–90. http://dx.doi.org/10.21314/JRMV.2014.117 doi: 10.21314/JRMV.2014.117
![]() |
[13] | Mendenhall W, Beaver RJ, Beaver BM (2008) Introduction to probability and statistics. Cengage Learning, 13th edition. |
[14] | Rolski T, Schmidli H, Schmidt V, et al. (1999) Stochastic Processes for Insurance and Finance. Wiley Series in Probability and Statistics. John Wiley Sons. |
[15] | Scandizzo S (2016) The validation of risk models: A handbook for practitioners. Springer. https://doi.org/10.1057/9781137436962 |
[16] |
Tasche D (2013a) The art of probability-of-default curve calibration. J Credit Risk, 9: 63–103. https://doi.org/10.21314/JCR.2013.169 doi: 10.21314/JCR.2013.169
![]() |
[17] | Tasche D (2013b) The Law of Total Odds. Working paper. https://doi.org/10.48550/arXiv.1312.0365. |
[18] | Venables WN, Ripley BD (2002) Modern Applied Statistics with S. Springer, fourth edition. |
![]() |
![]() |
1. | Yunpeng Xiao, Wen Teng, Representations and cohomologies of modified λ-differential Hom-Lie algebras, 2024, 9, 2473-6988, 4309, 10.3934/math.2024213 | |
2. | Wen Teng, Fengshan Long, Yu Zhang, Cohomologies of modified λ-differential Lie triple systems and applications, 2023, 8, 2473-6988, 25079, 10.3934/math.20231280 |