Let {Xn,n≥1} be a sequence of m-dependent strictly stationary random variables in a sub-linear expectation (Ω,H,E). In this article, we give the definition of m-dependent sequence of random variables under sub-linear expectation spaces taking values in R. Then we establish moderate deviation principle for this kind of sequence which is strictly stationary. The results in this paper generalize the result that in the case of independent identically distributed samples. It provides a basis to discuss the moderate deviation principle for other types of dependent sequences.
Citation: Shuang Guo, Yong Zhang. Moderate deviation principle for m-dependent random variables under the sub-linear expectation[J]. AIMS Mathematics, 2022, 7(4): 5943-5956. doi: 10.3934/math.2022331
[1] | Mingzhou Xu, Xuhang Kong . Note on complete convergence and complete moment convergence for negatively dependent random variables under sub-linear expectations. AIMS Mathematics, 2023, 8(4): 8504-8521. doi: 10.3934/math.2023428 |
[2] | He Dong, Xili Tan, Yong Zhang . Complete convergence and complete integration convergence for weighted sums of arrays of rowwise $ m $-END under sub-linear expectations space. AIMS Mathematics, 2023, 8(3): 6705-6724. doi: 10.3934/math.2023340 |
[3] | Mingzhou Xu, Kun Cheng, Wangke Yu . Complete convergence for weighted sums of negatively dependent random variables under sub-linear expectations. AIMS Mathematics, 2022, 7(11): 19998-20019. doi: 10.3934/math.20221094 |
[4] | Lunyi Liu, Qunying Wu . Complete integral convergence for weighted sums of negatively dependent random variables under sub-linear expectations. AIMS Mathematics, 2023, 8(9): 22319-22337. doi: 10.3934/math.20231138 |
[5] | Mingzhou Xu . On the complete moment convergence of moving average processes generated by negatively dependent random variables under sub-linear expectations. AIMS Mathematics, 2024, 9(2): 3369-3385. doi: 10.3934/math.2024165 |
[6] | Shuyan Li, Qunying Wu . Complete integration convergence for arrays of rowwise extended negatively dependent random variables under the sub-linear expectations. AIMS Mathematics, 2021, 6(11): 12166-12181. doi: 10.3934/math.2021706 |
[7] | Mingzhou Xu . Complete convergence of moving average processes produced by negatively dependent random variables under sub-linear expectations. AIMS Mathematics, 2023, 8(7): 17067-17080. doi: 10.3934/math.2023871 |
[8] | Mingzhou Xu . Complete convergence and complete moment convergence for maximal weighted sums of extended negatively dependent random variables under sub-linear expectations. AIMS Mathematics, 2023, 8(8): 19442-19460. doi: 10.3934/math.2023992 |
[9] | Qingfeng Wu, Xili Tan, Shuang Guo, Peiyu Sun . Strong law of large numbers for weighted sums of $ m $-widely acceptable random variables under sub-linear expectation space. AIMS Mathematics, 2024, 9(11): 29773-29805. doi: 10.3934/math.20241442 |
[10] | Baozhen Wang, Qunying Wu . Almost sure convergence for a class of dependent random variables under sub-linear expectations. AIMS Mathematics, 2024, 9(7): 17259-17275. doi: 10.3934/math.2024838 |
Let {Xn,n≥1} be a sequence of m-dependent strictly stationary random variables in a sub-linear expectation (Ω,H,E). In this article, we give the definition of m-dependent sequence of random variables under sub-linear expectation spaces taking values in R. Then we establish moderate deviation principle for this kind of sequence which is strictly stationary. The results in this paper generalize the result that in the case of independent identically distributed samples. It provides a basis to discuss the moderate deviation principle for other types of dependent sequences.
Moderate deviation principle (MDP) and large deviation principle (LDP) are applied to many aspects such as physics, finance and communication because of their typical significance. Especially in risk measurement and model uncertainty in statistics and finance. LDP and MDP can measure the asymptotic properties of the probability of rare event, particularly LDP and MDP for independent identically distributed (IID) random variables. Although LDP and MDP for the sequence of IID random variables are similar in form, there are essential differences between LDP and MDP. Generally speaking, LDP describes the ergodic phenomena that depend on the law of large numbers, while the MDP describes the asymptotic behavior of probabilities that between the law of large numbers and the central limit theorem. Specifically, the rate function of LDP depends on the distribution of random variables, while the rate function of MDP only depends on the limit density form of the central limit theorem of random variables, so MDP is more universal compared with LDP.
Petrov [23] constructed the law of the iterated logarithm and probabilities of moderate deviations of sums of dependent random variables. The main aim of Feng et al. [8] was to study self-normalized moderate deviations for transient random walk in random scenery under a much weaker condition than a finite moment-generating function of the scenery variables. Chen and Zhang [1] mainly discussed moderate deviations for the total population arising from a nearly unstable sub-critical Galton-Watson process with immigration. Xue [25] presented MDP for the paths of a certain class of density-dependent Markov chains. Miao et al. [18] obtained the conclusion that MDP for m-dependent random variables.
Peng [20] came up with a new notion of sub-linear expectation, referred to as G-expectation and the related G-normal distribution. Through the research of scholars, sub-linear expectation can be regarded as an extension of the classical linear expectation. There are two important nonlinear expectations, one is the capacity which was introduced by Choquet [4] and the other is sub-linear expectation spaces which was initiated by Peng [22]. Peng [22] also gave some basic concepts simultaneously. Furthermore, Denis et al. [5] presented the function spaces and capacity related to a sub-linear expectation: Application to G-Brownian motion paths. Liu and Zhang [17] established central limit theorem and invariance principle for linear processes generated by IID random variables under sub-linear expectation. Many related results have been investigated in the sub-linear expectation space. These results can be found in the work of Guo and Zhang [11], Lin and Feng [14], Liu and Zhang [16], and references therein. Recently, some researchers have already got LDP and MDP under the sub-linear expectation framework. Tan and Zong [24] obtained the LDP for random variables in sub-linear expectation spaces taking values in Rd, which do not have to satisfy the independent identical distributed condition. Chen and Feng [2] put forward LDP for negatively dependent random variables under sub-linear expectation, moreover they gained the upper bound of MDP. Associated with Peng's CLT, Gao and Xu [9] obtained LDP and MDP for independent random variables under sub-linear expectations. Zhou and Logachov [26] investigated MDP for a sequence of weak independent but not identically distributed random variables under sub-linear expectations. It can be regarded as an extension of the Theorem 3.1 and Corollary 3.2 in Gao and Xu [9] with d=1 and Yi=0.
The main aim of this paper is to prove MDP for m-dependent random variables under the sub-linear expectation. This theorem can be considered as a generalization of Gao and Xu's conclusion (in [9]). This paper is organized as follows: In Section 2, we introduce the framework of sub-linear expectation spaces and some basic settings. In Section 3, we give the main result and proofs of this paper.
First we review some definitions related symbols and properties on sub-linear expectation space. More detailed information are mentioned in Peng [20,21,22]. Let Ω be a complete separable metric space equipped with the the Borel σ-algebra B(Ω). We use the framework and notations of Peng [19,20,21,22].
Let (Ω,F) be a given measurable space. Let H be a linear space of real functions defined on Ω such that if (X1,X2,...,Xn)∈H, then φ(X1,X2,...,Xn)∈H for each φ∈Cl,Lip(Rn), where φ∈Cl,Lip(Rn) denotes the linear space of local Lipschitz continuous functions φ satisfying
|φ(x)−φ(y)|⩽C(1+|x|m+|y|m)|x−y|,∀x,y∈Rn, |
for some C>0,m∈N depending on φ. In this paper, the space H will be used as the space of random variables. Often a random variable X=(X1,X2,...,Xn) is called a n-dimensional random vector, denoted by X∈Hn.
Definition 2.1. A sub-linear expectation on H is a functional E:H→R satisfying the following properties: For any X,Y∈H, we have
(1) Monotonicity: X⩾Y implies E[X]⩾E[Y].
(2) Constant preserving: E[c]=c,∀c∈R.
(3) Sub-additivity: E[X+Y]⩽E[X]+E[Y].
(4) Positive homogeneity: E[λX]=λE[X],∀λ>0.
Here R=[−∞,+∞]. The triple (Ω,H,E) is called a sub-linear expectation space.
We give some useful properties of sub-linear expectation E, the proof can be found in Peng [22].
Remark 2.1. Properties (3) and (4) are called sub-linearity, they imply
(5) Convexity: E[αX+(1−α)Y]⩽αE[X]+(1−α)E[Y], for α∈[0,1].
If a nonlinear expectation E satisfies the convexity property, we call it a convex expectation.
Properties (2) and (3) imply
(6) Cash translatability: E[X+c]=E[X]+c,∀c∈R.
(7) From the definition, it is easily shown that E[X−Y]⩾E[X]−E[Y],∀X,Y∈H with E[Y] being finite.
Definition 2.2. (Identical distribution) (Peng [22]) Let X1 and X2 be two n-dimensional random vectors defined respectively in sub-linear expectation spaces (Ω1,H1,E1) and (Ω2,H2,E2). They are called identically distributed, denoted by X1d=X2, if
E1[φ(X1)]=E2[φ(X2)],∀φ∈Cl,Lip(Rn), |
whenever the sub-expectations are finite. A sequence of random variables {Xn,n⩾1} is said to be identically distributed if Xid=X1 for each i⩾1.
Definition 2.3. (Independence) (Chen and Feng [2]) Let (X1,…,Xn+1) be real random variables on (Ω,H). Xn+1 is said to be independent of (X1,…,Xn) under E[⋅], if for every non negative measurable function φi(⋅) on R with E[φi(Xi)]<∞,i=1,…,n+1, we have
E[n+1∏i=1φi(Xi)]=E[n∏i=1φi(Xi)]E[φn+1(Xn+1)]. |
{Xn}∞n=1 is said to be a sequence of independent random variables, if Xn+1 is independent of (X1,…,Xn) for all n∈N.
Definition 2.4. (IID random variables) A sequence of random variables {Xn,n⩾1} is said to be independent and identically distributed, if Xid=X1 and Xi+1 is independent to (X1,...,Xi) for each i⩾1.
Definition 2.5. (m-dependent random variables) (Li [13]) The sequence{Xi}∞i=1 is called m-dependent if there exits an integer m such that for every n and every j⩾m+1, (Xn+m+1,⋯,Xn+j) is independent from (X1,...,Xn). In particular, if m=0, {Xi}∞i=1 is called independent sequence.
Next we introduce a definition that generalizes the concept of strictly stationary sequence to the case of sequence in a sub-linear expectation space.
Definition 2.6. (Liu and Zhang [15]) {ξn,n∈N} is said to be a sequence of strictly stationary random variables on the (Ω,H,E), if for any function ϕn∈Cl,Lip(Rn):Rn→R, then
E[ϕn(ξ1,ξ2,...,ξn)]=E[ϕn(ξ1+k,ξ2+k,...,ξn+k)],∀n≥1,k∈N. |
In this paper, we always assume E is regular, i.e., for all {Xn,n⩾1}⊂H,
Xn(ω)↓0forallω∈Ω⟹limn→∞E(Xn)=0. |
Now we recall the definition of the upper expectation and corresponding capacity. For more details, refer to chapter 6 of Peng [22]. Let M be the set of all probability measures on Ω and denote P as relatively compact subset of M.
∙L0(Ω): the space of all B(Ω)-measurable real functions.
∙Bb(Ω): all bounded functions in L0(Ω).
∙Cb(Ω): all continuous functions in Bb(Ω).
We define the upper probability V
V(A)=supP∈PP(A),A∈B(Ω). |
Under this framework, V(⋅) is a Choquet capacity (cf. Denis and Martini [6], Dellacherie [7]) which satisfies the following properties.
Proposition 2.1. V(⋅) is a Choquet capacity, i.e.,
(1) 0⩽V(A)⩽1,∀A∈Ω.
(2) If A⊂B, then V(A)⩽V(B).
(3) If (An)∞n=1 is a sequence in B(Ω), then V(∞⋃n=1An)⩽∞∑n=1V(An).
(4) If (An)∞n=1 is an increasing sequence in B(Ω):An↑A=∞⋃n=1An, thenV(∞⋃n=1An)=limn→∞V(An).
We will consider the upper expectation E generated by P, for any X∈L0(Ω) such that EP[X] exists for each P∈P,
E[X]:=supP∈PEP[X],V(A)=E[IA],A∈B(Ω), |
and L1={X∈B(Ω);E[|X|]<∞}. The upper expectation E[⋅] is a sub-linear expectation which is introduced in Definition 2.1 on Bb(Ω) as well as on Cb(Ω).
In the end of this section, we recall two important lemmas which will be used in the following proof.
Lemma 2.1. (Chen et al. [3]) Let X,Y be real F-measurable random variables on upper expectation space (Ω,H,E).
(1) Hölder's inequality: For p,q>1 with 1p+1q=1, we have
E[|XY|]⩽(E[|X|p])1p⋅(E[|Y|q])1q. |
(2) Chebyshev's inequality: Let f(x)>0 be a non-decreasing nonnegative function on R, then for any x,
V(X⩾x)⩽E[f(X)]f(x). |
Some notations and definitions that will be used in this section are presented before we give the main results.
Definition 3.1. A function I:R→[0,∞) is called a rate function, if for all l⩾0, the set {x:I(x)⩽l} is a compact subset of R.
We will consider a sequence of m-independent random variables {Xi}∞i=1 in a sub-linear expectation space (Ω,H,E). Assume that bn is a sequence of positive real numbers satisfying limn→∞bn√n=∞,limn→∞bnn=0. Denote Sn=n∑i=1Xi,¯Sn=Snbn, for n∈N.
We adopt the framework and notations in Section 2 and recall the definition of LDP under sub-linear expectations with an upper probability V.
Definition 3.2. (Gao and Xu [10]) Let Ω be a topology space and R be a σ-algebra on Ω. Let (Vn,n⩾1) be a family of measurable maps from Ω into R and a(n), n⩾1 be a positive function satisfying a(n)→∞ as n→∞.
{V(Vn∈⋅),n⩾1} is said to satisfy LDP with speed a(n) and with rate function I(x) if for any measurable closed subset F⊂R,
lim supn→∞1a(n)logV(Vn∈F)⩽−infx∈FI(x), | (3.1) |
and for any measurable open set O⊂R,
lim infn→∞1a(n)logV(Vn∈O)⩾−infx∈OI(x), | (3.2) |
(3.1) and (3.2) are referred respectively to as upper bound of large deviations (ULD) and lower bound of large deviations (LLD).
If (3.1) and (3.2) are satisfied with a(n)=b2nn and Vn=Snbn, we say that Vn satisfies the LDP with rate function I(⋅) and speed nb2n, also say that Vn satisfies MDP.
Next we give the main result of this paper.
Theorem 3.1. Let {Xi}∞i=1 be a strictly stationary m-dependent random variable sequenceunder sub-linear expectation, E[X1]=E[−X1]=0, limn→∞1nE[n∑i=1Xi]2=σ2<∞, there existsδ>0, such that Eexp{δ|X1|}<∞. {V(¯Sn∈⋅),n⟶∞} is said to satisfy the MDP with a rate function I(x).
(1) For every closed set F⊂R such that
lim supn→∞nb2nlogV(¯Sn∈F)⩽−infx∈FI(x); | (3.3) |
(2) for every open set G⊂R such that
lim infn→∞nb2nlogV(¯Sn∈G)⩾−infx∈GI(x), | (3.4) |
where I(x):=supy∈R{xy−y2σ22}=x22σ2.
Before starting our proof, we need to mention the following lemmas which are useful to prove the above results.
Lemma 3.1. (Liu and Zhang [15]) Suppose that {θi,1⩽i⩽n} is a sequence of R-valued random variables on (Ω,H,E) and ζi∈[0,1], 1⩽i⩽n, such that n∑i=1ζi=1.Then
logEexp{n∑i=1ζiθi}⩽n∑i=1ζilogEexp{θi}. | (3.5) |
Lemma 3.2. Under the assumption of Theorem 3.1, for any x∈R, we have
limn→∞nb2nlogEexp(xbnnSn)=σ2x22. | (3.6) |
Proof. Fix K⩾m+1,n⩾K+m, define l=[nK+m]. Note
εt=K∑i=1Xt(K+m)+i,ηt=m∑i=1Xt(K+m)+K+i,t=0,1,⋯,l−1. |
We have defined that {Xi,i⩾1} is zero-mean m-dependent random variables, therefore {εt,t⩾0} and {ηt,t⩾0} are both independent identically distributed random variables, in addition Eε1=0, Eη1=0. Thus, based on the definition of Sn, we have
Sn=l−1∑t=0εt+l−1∑t=0ηt+n∑i=l(K+m)+1Xi:=I1+I2+I3, | (3.7) |
when p1>1,p2>1, choose q1>1,q2>1, such that 1p1+1q1=1,1p2+1q2=1. By the means of (3.7) and (1) in Lemma 2.1 (Hölder inequality), for any x∈R, that
nb2nlogEexp{xbnnSn}⩽1q1⋅nb2nlogEexp{xq1bnnI3}+1p1⋅nb2nlogEexp{xp1bnn(I1+I2)}⩽1q1⋅nb2nlogEexp{xq1bnnI3}+1p1p2⋅nb2nlogEexp{xp1p2bnnI1}+1p1q2⋅nb2nlogEexp{xp1q2bnnI2}:=J1+J2+J3. | (3.8) |
Above all, we deal with J1. By using the definition of I3, Lemma 3.1 and the strictly stationary of {Xi}∞i=1, it's obvious that
lim supn→∞J1⩽limn→∞1q1⋅nb2nlogEexp{|x|q1bnn(l+1)(K+m)∑i=l(K+m)+1|Xi|}=lim supn→∞1q1⋅nb2nlogEexp{|x|q1bnn(K+m)K+m∑i=1|Xi|K+m}⩽lim supn→∞1q1⋅nb2nK+m∑i=11K+mlogEexp{|x|q1bnn(K+m)|Xi|}=lim supn→∞1q1⋅nb2nlogEexp{|x|q1bnn(K+m)|X1|}⩽lim supn→∞1q1⋅nb2nlogEexp{δ|X1|}=0, | (3.9) |
when n is sufficient big, then |x|q1bnn(K+m)⩽δ.
Next is to prove J2. According to Lemma 2.1 in Zhou and Logachov [26], we know that
limn→∞n2b2nlogEexp{ybnnε1}=σ21y22,∀y∈R, | (3.10) |
where σ21=Eε21. On the one hand, by the definition of l, thatlimn→∞ln=1K+m. On the other hand, {εi} isindependent identically distributed and {Xi} is strictly stationary, that
limn→∞J2=1p1p2limn→∞nb2nlogEexp{xp1p2bnnl−1∑t=0εt}=1p1p2limn→∞lnb2nlogEexp{xp1p2bnnε0}=1p1p2⋅1K+mlimn→∞n2b2nlogEexp{xp1p2bnnε0}=p1p2⋅1K+m⋅σ21x22=p1p2⋅1K+m⋅x22E[K∑i=1Xi]2=p1p2⋅KK+m⋅x22⋅1KE[K∑i=1Xi]2. |
Firstly we order K→∞, then let p1→1,p2→1, that
limp1→1,p2→1limK→∞limn→∞J2=limp1→1,p2→1limK→∞p1p2x22⋅KK+m⋅1KE[K∑i=1Xi]2=limp1→1,p2→1p1p2x22⋅σ2=σ2x22. | (3.11) |
Now we prove J3. Because {ηi} is also independent identically distributed, so we use the same method with J2, such that
limn→∞J3=1p1q2limn→∞nb2nlogEexp{xp1q2bnnl−1∑t=0ηt}=1p1q2limn→∞lnb2nlogEexp{xp1q2bnnη0}=1p1q2⋅1K+mlimn→∞n2b2nlogEexp{xp1q2bnnη0}=p1q2K+m⋅x2Eη212=p1q2K+m⋅x22E[m∑i=1Xi]2. |
Let K→∞, so we can obtain
limK→∞limn→∞J3=0. | (3.12) |
Combining with (3.9), (3.11) and (3.12), we can get
lim supn→∞nb2nlogEexp{xbnnSn}⩽σ2x22. | (3.13) |
According to (1) in Lemma 2.1 (Hölder inequality), that
logEexp{xbnp1p2nI1}=logEexp{xbnp1p2n(I1+I2−I2)}⩽1p1logEexp{bnp2n(I1+I2)}+1q1logEexp{−xq1bnp1p2nI2}=1p1logEexp{xbnp2n(I1+I2+I3−I3)}+1q1logEexp{−xq1bnp1p2nI2}⩽1p1p2logEexp{xbnn(I1+I2+I3)}+1p1q2logEexp{−xq2bnp2nI3}+1q1logEexp{−xq1bnp1p2nI2}. | (3.14) |
Similar to (3.9), that
lim supn→∞1p1q2⋅nb2nlogEexp{−xq2bnp2nI3}⩽0. | (3.15) |
Similar to (3.11), that
limp1→1,p2→1limK→∞limn→∞nb2nlogEexp{xbnp1p2nI1}=σ2x22. | (3.16) |
Similar to (3.12), that
limK→∞limn→∞1q1⋅nb2nlogEexp{−xq1bnp1p2nI2}=0. | (3.17) |
Combining with (3.14)–(3.17), we conclude that
σ2x22=limp1→1,p2→1limK→∞limn→∞nb2nlogEexp{xbnp1p2nI1}⩽limp1→1,p2→1lim infn→∞1p1p2⋅nb2nlogEexp{xbnnSn}+lim supn→∞1p1q2⋅nb2nlogEexp{−xq2bnp2nI3}+limK→∞limn→∞1q1⋅nb2nlogEexp{−xq1bnp1p2nI2}=lim infn→∞nb2nlogEexp{xbnnSn}, | (3.18) |
and together with (3.13) and (3.18), then follows (3.6). Thus we finish the proof.
Proof of Theorem 3.1
Proof. We divide the proof into two steps in this section.
Step 1. Let F be any given closed set. It's obvious we can get the result when F=∅. So we suppose that F≠∅. Note
x−:=sup{x∈F:x<0}⩽0,x+:=inf{x∈F:x⩾0}⩾0, |
thus F⊆(−∞,x−]∪[x+,+∞). Let x−=−∞, if F∩(−∞,0]=∅ and x+=+∞, if F∩[0,+∞)=∅. Then, if F≠∅, there exist a finite x− or x+. Hence it is sufficient to see that
logV(¯Sn∈F)=logEI{¯Sn∈F}⩽log(EI{¯Sn∈(−∞,x−]}+EI{¯Sn∈[x+,+∞)})⩽log(2max(EI{¯Sn∈(−∞,x−]},EI{¯Sn∈[x+,+∞)})). | (3.19) |
Now we estimate EI{¯Sn∈(−∞,x−]} primarily. Let f(x)=ex, for all λ>0, by (2) in Lemma 2.1 (Chebyshev's inequality) we can get
EI{¯Sn∈(−∞,x−]}=EI{−λbnnSn⩾−λb2nnx−}⩽Ee−λbnnSne−λb2nnx−, |
from Lemma 3.2, that,
nb2nlogV{¯Sn∈(−∞,x−]}⩽nb2nlogEe−λbnnSne−λb2nnx−=nb2nlogEe−λbnnSn+λx−, |
therefore
lim supn→∞nb2nlogV{¯Sn∈(−∞,x−]}⩽lim supn→∞nb2nlogEexp{−λbnnSn}+λx−=σ2λ22+λx−. | (3.20) |
Here we choose λ=−x−σ2 in (3.20) thus
lim supn→∞nb2nlogV{¯Sn∈(−∞,x−]}⩽−x2−2σ2. | (3.21) |
Similarly, it is obvious that
lim supn→∞nb2nlogV{¯Sn∈[x+,+∞)}⩽−x2+2σ2. | (3.22) |
Hence, from (3.19),(3.21) and (3.22) it is sufficient to show that
lim supn→∞nb2nln(V(¯Sn∈F))⩽−min(x2−2σ2,x2+2σ2)⩽−infx∈FI(x). |
Finally, infx∈FI(x)⩽min(x2−2σ2,x2+2σ2) is remain to prove. In fact, under the condition we have given that I(x):=supy∈R{xy−y2σ22}=x22σ2, it yields
infx∈FI(x)⩽min(I(x−),I(x+))=min(x2−2σ2,x2+2σ2). |
Therefore, we have finish the proof of the inequality (3.3) in Theorem 3.1.
Step 2. Next we have to prove that for any open set G inequality (3.4) holds with I(x) defined in Theorem 3.1. Let G be any open set. It is obvious we can gain the result when G=∅. In order to get the following assumption we choose G≠∅. It's apparent that for any h⩾0, the set
Th:={x:I(x)⩽h}, |
is compact. According to G≠∅, then G∩ThG≠∅ as hG>0.
Due to G is an open set, for any μ>0 there exists x∈G∩ThG, we have
infy∈GI(y)⩾I(x)−μ. | (3.23) |
For any ξ>0 denote
x(ξ):=(x−ξ,x+ξ). |
For any λ∈R and n∈N, and denote O(λ,n):=logEeλbnnSn. Hence we can obtain
I{¯Sn∈x(ξ)}eλbnnSn=I{¯Sn∈x(ξ)}eλb2nn¯Sn⩾I{¯Sn∈x(ξ)}eb2nn(λx−|λ|ξ). | (3.24) |
By (3.24) and (7) in Remark 2.1, for any λ∈R and sufficiently small ξ>0, it yields
logV(¯Sn∈G)⩾logV(¯Sn∈x(ξ))=logEI{¯Sn∈x(ξ)}⩾logE(I{¯Sn∈x(ξ)}eλbnnSn−O(λ,n)e−λb2nnx+O(λ,n)e−|λ|b2nnξ)=−λb2nnx+O(λ,n)−|λ|b2nnξ+logE{(1−I{¯Sn∉x(ξ)})eλbnnSn−O(λ,n)}⩾−λb2nnx+O(λ,n)−|λ|b2nnξ+log(1−EI{¯Sn∉x(ξ)}eλbnnSn−O(λ,n)). | (3.25) |
For all γ>0, we obtain that
EI{¯Sn∉x(ξ)}eλbnnSn−O(λ,n)⩽EI{¯Sn⩾x+ξ}eλbnnSn−O(λ,n)+EI{¯Sn⩽x−ξ}eλbnnSn−O(λ,n)⩽Eeγbnnσ2SneλbnnSneγb2n(x+ξ)σ2neO(λ,n)+Ee−γbnnσ2SneλbnnSneγb2n(ξ−x)σ2neO(λ,n)=Ee(γσ2+λ)bnnSneγb2n(x+ξ)σ2neO(λ,n)+Ee(λ−γσ2)bnnSneγb2n(ξ−x)σ2neO(λ,n)=H1+H2. |
The following point is to proof H1 and H2, respectively. From Lemma 3.2 we know that
limn→∞nb2nlogEexp{λbnnSn}=λ2σ22, |
there exist N, such that n>N, for any ϵ∈R, we have
EeλbnnSn⩽eb2nnλ2σ22(1+ϵ). |
For n large enough, thus
H1=Ee(γσ2+λ)bnnSneb2nnγx+γξσ2eO(λ,n)⩽eb2nn((γ/σ2+λ)22σ2(1+ϵ))eb2nn(γx+γξσ2+λ22σ2(1+ϵ))=exp{b2nn[(γ22σ2+γλ)(1+ϵ)−γx+γξσ2]}, |
by the choice of γ=ξ, let ϵ→0 thus
H1⩽exp{b2nn[−ξ22σ2+ξ(λ−xσ2)]}. |
Similarly
H2⩽exp{b2nn[−ξ22σ2+ξ(xσ2−λ)]}. |
It can be concluded that
EI{¯Sn∉x(ξ)}eλbnnSn−O(λ,n)⩽2exp{b2nn[−ξ22σ2+ξ|xσ2−λ|]}, |
we choose λ=xσ2, thus
nb2nlog(1−EI{¯Sn∉x(ξ)}eλbnnSn−O(λ,n))⩾nb2nlog{1−(H1+H2)}⩾nb2nlog{1−2exp{b2nn[−ξ22σ2+ξ|xσ2−λ|]}}→0. | (3.26) |
Combining (3.25),(3.26) and Lemma 3.2, we get
lim infn→∞nb2nlogV(¯Sn∈G)⩾−x22σ2−|x|σ2ξ, |
let ξ→0, since x∈ThG and the set ThG does not depend on ξ, so we obtain
lim infn→∞nb2nlogV(¯Sn∈G)⩾−x22σ2. | (3.27) |
By (3.27), for any μ>0
lim infn→∞nb2nlogV(¯Sn∈G)⩾−x22σ2=−I(x)⩾−infx∈GI(x)−μ. | (3.28) |
Consequently, when μ→0 the result is deduced from (3.28). Combining Step 1 and Step 2, the proof of Theorem 3.1 is obtained.
In this paper, the author establish MDP for m-dependent random variables under the sub-linear expectation. The results extend Gao and Xu's conclusion (in [9]) from independent random variables to m-dependent random variables under the sub-linear expectations. The research about MDP for dependent sequences is a new trend in probability and statistics, one can refer to [9,18,25,26] and references therein for details.
The paper was supported by National Natural Science Foundation of China (Grant No. 11771178, 12171198); the Science and Technology Development Program of Jilin Province (Grant No. 20210101467JC) and Science and Technology Program of Jilin Educational Department during the "13th Five-Year" Plan Period (Grant No. JJKH20200951KJ) and Fundamental Research Funds for the Central Universities.
The authors declare no conflict of interest in this paper.
[1] |
H. T. Chen, Y. Zhang, Moderate deviations for the total population arising from a nearly unstable sub-critical Galton-Watson process with immigration, Commun. Stat.-Theor. M., 50 (2021), 432–445. http://dx.doi.org/10.1080/03610926.2019.1635703 doi: 10.1080/03610926.2019.1635703
![]() |
[2] |
Z. J. Chen, X. W. Feng, Large deviation for negatively dependent random variables under sublinear expectation, Commun. Stat.-Theor. M., 45 (2016), 400–412. http://dx.doi.org/10.1080/03610926.2015.1006067 doi: 10.1080/03610926.2015.1006067
![]() |
[3] |
Z. J. Chen, P. Y. Wu, B. M. Li, A strong law of large numbers for non-additive probabilities, Int. J. Approx. Reason., 54 (2013), 365–377. http://dx.doi.org/10.1016/j.ijar.2012.06.002 doi: 10.1016/j.ijar.2012.06.002
![]() |
[4] |
G. Choquet, Theory of capacities, Ann. I. Fourier, 5 (1953), 131–295. http://dx.doi.org/10.5802/aif.53 doi: 10.5802/aif.53
![]() |
[5] |
L. Denis, M. Hu, S. G. Peng, Function spaces and capacity related to a sublinear expecation: Application to G-Brownian motion paths, Potential Anal., 34 (2011), 139–161. http://dx.doi.org/10.1007/s11118-010-9185-x doi: 10.1007/s11118-010-9185-x
![]() |
[6] |
L. Denis, C. Martini, A theoretical framework for the pricing of contingent claims in the presence of model uncertainty, Ann. Appl. Probab., 16 (2006), 827–852. http://dx.doi.org/10.1214/105051606000000169 doi: 10.1214/105051606000000169
![]() |
[7] | C. Dellacherie, Capacités et processus stochastiques, Springer-Verlag, New York, 1972. |
[8] |
X. W. Feng, Q. M. Shao, O. Zeitouni, Self-normalized moderate deviations for random walk in random scenery, J. Theor. Probab., 34 (2021), 103–124. http://dx.doi.org/10.1007/s10959-019-00965-2 doi: 10.1007/s10959-019-00965-2
![]() |
[9] |
F. Q. Gao, M. Z. Xu, Large deviations and moderate deviations for independent random variables under sublinear expectations, Sci. Sin. Math., 41 (2011), 337–352. http://dx.doi.org/10.1360/012009-879 doi: 10.1360/012009-879
![]() |
[10] |
F. Q. Gao, M. Z. Xu, Relative entropy and large deviations under sublinear expectations, Acta Math. Sci. Ser. B, 32 (2012), 1826–1834. http://dx.doi.org/10.1016/S0252-9602(12)60143-X doi: 10.1016/S0252-9602(12)60143-X
![]() |
[11] | S. Guo, Y. Zhang, Central limit theorem for linear processes generated by m-dependent random variables under the sub-linear expectation, Commun. Stat.-Theor. M., 2022, In press. |
[12] |
Z. C. Hu, L. Zhou, Multi-dimensional central limit theorems and laws of large numbers under sublinear expectations, Acta Math. Sin., 31 (2015), 305–318. http://dx.doi.org/ 10.1007/s10114-015-3212-1 doi: 10.1007/s10114-015-3212-1
![]() |
[13] |
X. P. Li, A central limit theorem for m-dependent random variables under sublinear expectations, Acta Math. Appl. Sin.-E., 31 (2015), 435–444. http://dx.doi.org/ 10.1007/s10255-015-0477-1 doi: 10.1007/s10255-015-0477-1
![]() |
[14] |
Y. W. Lin, X. W. Feng, Complete convergence and strong law of large numbers for arrays of random variables under sublinear expectations, Commun. Stat.-Theor. M., 49 (2020), 5866–5882. http://dx.doi.org/10.1080/03610926.2019.1625924 doi: 10.1080/03610926.2019.1625924
![]() |
[15] |
W. Liu, Y. Zhang, Large deviation principle for linear processes generated by real stationary sequences under the sub-linear expectation, Commun. Stat.-Theor. M., 2021. http://dx.doi.org/10.1080/03610926.2021.2018462 doi: 10.1080/03610926.2021.2018462
![]() |
[16] |
W. Liu, Y. Zhang, The law of the iterated logarithm for linear processes generated by stationary independent random variables under the sub-linear expectation, Entropy, 23 (2021), 1313. http://dx.doi.org/10.3390/e23101313 doi: 10.3390/e23101313
![]() |
[17] |
W. Liu, Y. Zhang, Central limit theorem for linear processes generated by IID random variables under the sub-linear expectation, Appl. Math. Ser. B, 36 (2021), 243–255. http://dx.doi.org/10.1007/s11766-021-3882-7 doi: 10.1007/s11766-021-3882-7
![]() |
[18] |
Y. Miao, J. N. Zhu, J. Y. Mu, Moderate deviation principle for m-dependent random variables, Lith. Math. J., 58 (2018), 54–68. http://dx.doi.org/10.1007/s10986-018-9382-7 doi: 10.1007/s10986-018-9382-7
![]() |
[19] |
S. G. Peng, Monotonic limit theorem of BSDE and nonlinear decomposition theorem of Doob-Meyers type, Probab. Theory Rel., 113 (1999), 473–499. http://dx.doi.org/10.1007/s004400050214 doi: 10.1007/s004400050214
![]() |
[20] |
S. G. Peng, G-Expectation, G-Brownian motion and related stochastic calculus of Ito's type, Stoch. Anal. Appl., 2 (2006), 541–567. http://dx.doi.org/10.1007/978-3-540-70847-6-25 doi: 10.1007/978-3-540-70847-6-25
![]() |
[21] |
S. G. Peng, Survey on normal distributions, central limit theorem, Brownian motion and the related stochastic calculus under sublinear expectations, Sci. China Math., 52 (2009), 1391–1411. http://dx.doi.org/10.1007/s11425-009-0121-8 doi: 10.1007/s11425-009-0121-8
![]() |
[22] | S. G. Peng, Nonlinear expectations and stochastic calculus under uncertainty, Springer, Berlin, Heidelberg, 2019. http://dx.doi.org/10.1007/978-3-662-59903-7 |
[23] |
V. V. Petrov, The law of the iterated logarithm and probabilities of moderate deviations of sums of dependent random variables, J. Math. Sci., 251 (2020), 128–130. http://dx.doi.org/10.1007/s10958-020-05072-w doi: 10.1007/s10958-020-05072-w
![]() |
[24] |
Y. Z. Tan, G. F. Zong, Large deviation principle for random variables under sublinear expectations on Rd, J. Math. Anal. Appl., 488 (2020), 1–14. http://dx.doi.org/10.1016/j.jmaa.2020.124110 doi: 10.1016/j.jmaa.2020.124110
![]() |
[25] |
X. F. Xue, Moderate deviations of density-dependent Markov chains, Stoch. Proc. Appl., 140 (2021), 49–80. http://dx.doi.org/10.1016/j.spa.2021.06.005 doi: 10.1016/j.spa.2021.06.005
![]() |
[26] |
Q. Q. Zhou, A. V. Logachov, Moderate deviations principle for independent random variables under sublinear expectations, Sib. Elektron. Math. Re., 18 (2021), 817–826. http://dx.doi.org/10.33048/semi.2021.18.060 doi: 10.33048/semi.2021.18.060
![]() |
1. | Li Wang, Qunying Wu, Almost sure convergence theorems for arrays under sub-linear expectations, 2022, 7, 2473-6988, 17767, 10.3934/math.2022978 | |
2. | He Dong, Xili Tan, Yong Zhang, Complete convergence and complete integration convergence for weighted sums of arrays of rowwise m-END under sub-linear expectations space, 2023, 8, 2473-6988, 6705, 10.3934/math.2023340 | |
3. | Yansong Bai, Yong Zhang, Congmin Liu, Zhiming Wang, Moderate deviation principle for different types of classical likelihood ratio tests, 2022, 0361-0926, 1, 10.1080/03610926.2022.2144376 | |
4. | Peiyu Sun, Dehui Wang, Xue Ding, Xili Tan, Yong Zhang, Moderate Deviation Principle for Linear Processes Generated by Dependent Sequences under Sub-Linear Expectation, 2023, 12, 2075-1680, 781, 10.3390/axioms12080781 | |
5. | TAN Xili, DONG He, SUN Peiyu, ZHANG Yong, Almost Sure Convergence of Weighted Sums for m-END Sequences under Sub-linear Expectations, 2024, 1, 3006-0656, 10.59782/sidr.v1i1.26 | |
6. | Xue Ding, Yong Zhang, Chover’s law of the iterated logarithm for weighted sums under sub-linear expectations, 2024, 53, 0361-0926, 6055, 10.1080/03610926.2023.2239399 | |
7. | Peiyu Sun, Dehui Wang, Xili Tan, Equivalent Conditions of Complete p-th Moment Convergence for Weighted Sum of ND Random Variables under Sublinear Expectation Space, 2023, 11, 2227-7390, 3494, 10.3390/math11163494 | |
8. | Xue Ding, Complete convergence for moving average process generated by extended negatively dependent random variables under sub-linear expectations, 2024, 53, 0361-0926, 8166, 10.1080/03610926.2023.2279924 | |
9. | Xue Ding, Yong Zhang, Precise asymptotics for maxima of partial sums under sub-linear expectation, 2024, 0361-0926, 1, 10.1080/03610926.2024.2435585 |