Loading [MathJax]/jax/output/SVG/jax.js
Research article

On the construction of recurrent fractal interpolation functions using Geraghty contractions

  • Received: 21 July 2023 Revised: 22 September 2023 Accepted: 07 October 2023 Published: 26 October 2023
  • The recurrent iterated function systems (RIFS) were first introduced by Barnsley and Demko and generalized the usual iterated function systems (IFS). This new method allowed the construction of more general sets, which do not have to exhibit the strict self similarity of the IFS case and, in particular, the construction of recurrent fractal interpolation functions (RFIF). Given a data set {(xn,yn)I×R,n=0,1,,N} where I=[x0,xN], we ensured that attractors of RIFS constructed using Geraghty contractions were graphs of some continuous functions which interpolated the given data. Our approach goes beyond the classical framework and provided a wide variety of systems for different approximations problems and, thus, gives more flexibility and applicability of the fractal interpolation method. As an application, we studied the error rates of time series related to the vaccination of COVID-19 using RFIF, and we compared them with the obtained results on the FIF.

    Citation: Najmeddine Attia, Hajer Jebali. On the construction of recurrent fractal interpolation functions using Geraghty contractions[J]. Electronic Research Archive, 2023, 31(11): 6866-6880. doi: 10.3934/era.2023347

    Related Papers:

    [1] Najmeddine Attia, Mohamed balegh, Rim Amami, Rimah Amami . On the Fractal interpolation functions associated with Matkowski contractions. Electronic Research Archive, 2023, 31(8): 4652-4668. doi: 10.3934/era.2023238
    [2] Yantong Guo, Quansheng Wu, Xiaofeng Wang . An extension of high-order Kou's method for solving nonlinear systems and its stability analysis. Electronic Research Archive, 2025, 33(3): 1566-1588. doi: 10.3934/era.2025074
    [3] Jin Li, Yongling Cheng . Barycentric rational interpolation method for solving time-dependent fractional convection-diffusion equation. Electronic Research Archive, 2023, 31(7): 4034-4056. doi: 10.3934/era.2023205
    [4] Jin Li, Yongling Cheng . Barycentric rational interpolation method for solving fractional cable equation. Electronic Research Archive, 2023, 31(6): 3649-3665. doi: 10.3934/era.2023185
    [5] Tingting Du, Zhengang Wu . On the reciprocal sums of products of $ m $th-order linear recurrence sequences. Electronic Research Archive, 2023, 31(9): 5766-5779. doi: 10.3934/era.2023293
    [6] Pinchao Meng, Xinyu Wang, Weishi Yin . ODE-RU: a dynamical system view on recurrent neural networks. Electronic Research Archive, 2022, 30(1): 257-271. doi: 10.3934/era.2022014
    [7] David Cheban, Zhenxin Liu . Averaging principle on infinite intervals for stochastic ordinary differential equations. Electronic Research Archive, 2021, 29(4): 2791-2817. doi: 10.3934/era.2021014
    [8] Muajebah Hidan, Ahmed Bakhet, Hala Abd-Elmageed, Mohamed Abdalla . Matrix-Valued hypergeometric Appell-Type polynomials. Electronic Research Archive, 2022, 30(8): 2964-2980. doi: 10.3934/era.2022150
    [9] Li Du, Xiaoqin Yuan . The minimality of biharmonic hypersurfaces in pseudo-Euclidean spaces. Electronic Research Archive, 2023, 31(3): 1587-1595. doi: 10.3934/era.2023081
    [10] Yao Sun, Lijuan He, Bo Chen . Application of neural networks to inverse elastic scattering problems with near-field measurements. Electronic Research Archive, 2023, 31(11): 7000-7020. doi: 10.3934/era.2023355
  • The recurrent iterated function systems (RIFS) were first introduced by Barnsley and Demko and generalized the usual iterated function systems (IFS). This new method allowed the construction of more general sets, which do not have to exhibit the strict self similarity of the IFS case and, in particular, the construction of recurrent fractal interpolation functions (RFIF). Given a data set {(xn,yn)I×R,n=0,1,,N} where I=[x0,xN], we ensured that attractors of RIFS constructed using Geraghty contractions were graphs of some continuous functions which interpolated the given data. Our approach goes beyond the classical framework and provided a wide variety of systems for different approximations problems and, thus, gives more flexibility and applicability of the fractal interpolation method. As an application, we studied the error rates of time series related to the vaccination of COVID-19 using RFIF, and we compared them with the obtained results on the FIF.



    In 1986, Barnsley [1] introduced the concept of the fractal interpolation functions (FIFs) as continuous functions interpolating a given set of data points. Following this, an iterative process for interpolating data that sat on a discrete closed subset of Rn was introduced by Dubuc [2,3] and a general construction was given by Massopust [4]. This construction is based on the Read-Bajraktarevic operator for the FIFs. More precisely, FIFs are generated by an iterated function system (IFS) and arise as fixed points of the Read-Bajraktarevic operator defined on suitable function space. Several authors were interested in the study of many significant properties of FIFs including calculus, dimension, smoothness, stability, perturbation error, and more ([5,6,7]). The problem of the existence of FIF returns to the study of the existence and uniqueness of some fixed points on the fractal space, and the most widely studied FIFs are based on the Banach fixed point theorem. Recently, many researchers have studied the existence of FIFs by using different well-known results obtained in the fixed point theory, like Rakotch and Matkowski's theorems [9,10,11], Geraghty's theorem [12] and so on.

    Fractal interpolation has found applications in the field of image segmentation, image analysis, image compression and signal processing ([13,14]). In particular, uniformly sampled signal is required for the discrete wavelet transform computation when we consider the multifractal analysis to study the heart rate variability [15]. In order to gain more flexibility in natural-shape generation or in image compression, a new class of IFSs were introduced and were called RIFS [16,17]. The RIFS use elements of stochastic processes, which can provide more flexible methods since they involve local IFS [18,19]. It induces recurrent fractal interpolation functions (RFIFs) which main advantage is that they are not self-similar and they reflect only local similarities in contrast to the FIFs induced by the standard IFSs. Some properties of RFIFs, including statistical properties [20,21] and box dimension [22], have been studied (see also [17,23]). However, only RFIF, whose existence is ensured by the Banach fixed point theorem, has been considered.

    In this paper, we aim to construct non-affine RFIFs using Geraghty contractions. Therefore, we present in the next section some results on RIFS and we give new conditions on the existence of the attractor of a new class of RIFS. These results will be explored in the third paragraph to define a new class of RFIF using Gheraghty contractions. Section four will be devoted to the study of the normalized root mean square errors of time series related to the vaccination of COVID-19 using RFIF, and we compare them with the obtained results on the FIF.

    Let (X,d) be a complete, locally compact metric space and wj:XX, for j=1,,n are self-mappings on X. We define P=(pij) to be an irreducible n×n row stochastic matrix associated to these maps, i.e.,

    nj=1pij=1, for all i=1,,n and pij0, for all 1i,jn, (2.1)

    and for any 1i,jn, there exists a sequence i1, i2,,ik with i1=i and ik=j such that

    pi1i2pi2i3pik1ik>0. (2.2)

    A RIFS consists of an IFS (X,wi,i=1,,n) together with a matrix P as defined above. They have their origins in the work of Barnsley and Demko [24] and have been developed by Barnsley et al. [17]. In fact, they are a natural generalization of the IFS that permits them to construct fractals as fixed points of some operators on complete metric spaces [16,25].

    A RIFS is, in essence, a Markov chain with n states. Given an initial set that is transformed through the mapping wi, we should say that the chain is in state i. The value pij is the transition probability from state i to state j. That is, assuming that the mapping wi has been employed at the previous iteration, the probability of selecting the map wj for the next iteration is given by pij. The motivation underlying the necessary properties, that matrix (pij) should satisfy becomes apparent as a consequence of the Markov chain rationale. The first condition given by (2.1) ensures that the chain must be stochastic, that is, at every state the transition probabilities to any other state should sum up to one. The second one (2.2) means that the chain is irreducible. Namely, that we may transfer from any state of the chain to any other state. Therefore, we can be sure that, given sufficient time, all the mappings of the RIFS will be used. Otherwise, if P is reducible, it is possible that some of the mappings of the RIFS will be rendered useless.

    Remark 1. An IFS with probabilities {X,w1,,wn,p1,,pn} [16] provides a simple example of an RIFS {X,w1,,wn,(pij)1i,jn} when pij=pj, for all i,j{1,,n}. In addition, if p1=p2==pn:=p, then the RIFS reduces to the usual IFS denoted simply by (X,w,p). Therefore, RIFSs provide more flexibility so that their attractors need not exhibit the self-similarity or self-tiling properties characteristic of IFS attractors.

    Let H be the set of all nonempty compact subsets of X. We define the Hausdorff metric h by

    h(A,B)=max{d(A,B),d(B,A)},A,BH,

    with

    d(A,B)=supxAinfyBd(x,y) and d(B,A)=supxBinfyAd(x,y).

    Then [16], the space (H,h) is complete, and compact whenever X is compact. Let {X,w1,,wn,(pij)1i,jn} be a RIFS and let Wij:HH be defined as

    Wij(A)={wi(A)if pji>0.

    Now, consider the space Hn=H××H endowed with the metric ˜h defined as

    ˜h((A1,,An),(B1,,Bn))=max1inh(Ai,Bi)

    and let W:HnHn be given by

    W(A1A2An)=(W11W12W1nW21W22W2nWn1Wn2Wnn)(A1A2An)=(W11(A1)W12(A2)W1n(An)W21(A1)W22(A2)W2n(An)Wn1(A1)Wn2(A2)Wnn(An)).

    Thus,

    W(A1A2An)=(jI(1)w1(Aj)jI(2)w2(Aj)jI(n)wn(Aj)), (2.3)

    where I(i)={1jn pji>0}, for i=1,,n.

    Example 1. We consider I={[0,1]2,wi,1=1,,4} the RIFS equipped with the mappings

    w1(xy)=(1/2001/2)(xy)+(00),w2(xy)=(1/2001/2)(xy)+(050)w3(xy)=(1/2001/2)(xy)+(500),w4(xy)=(1/2001/2)(xy)+(5050). (2.4)

    If we consider the transition matrix

    P=(1/201/201/31/301/31/21/20001/201/2),

    then, the function W is defined by

    W=(w1w1w1w2w2w2w3w4w4).

    Now, let A=(A1,A2,A3,A4)TH4 be an initial point and assume that Ai was generated by applying the mapping wi to an unknown set for each i=1,2,3,4, then W(A) is computed as

    W(A1A2A3A4)=(w1(A1)w1(A2)w1(A3)w2(A2)w2(A3)w2(A4)w3(A1)w4(A2)w4(A4)).

    Therefore, only the maps wj with positive transition probabilities, i.e., pij>0 can be applied to the set Ai. In particular, only the map w3 is applied to A1.

    Although the Banach fixed point theorem is the most useful tool for the existence (and uniqueness) of attractors of IFS, there are numerous results in the literature related to some generalizations of the Banach principle that give sufficient conditions for the existence of a contractive fixed point (see for example [26,27,28]).

    Definition 1. 1) Let (X,ρ) be a metric space. If for some function φ:(0,)(0,) and a self-map f, we have :

     x,yX,ρ(f(x),f(y))φ(ρ(x,y)),

    then we say that f is a φ-contraction.

    2) If f is a φ-contraction for some function φ:(0,)(0,) such that for any t>0, α(t)=φ(t)t<1 and the function tφ(t)t is nonincreasing (so-called Geraghty I) (or nondecreasing (Geraghty II), or continuous (Geraghty III)), then we call such a function a Geraghty contraction.

    It is easy to see that each Banach contraction is a Geraghty contraction. It suffices to consider the function φ(t)=αt, for some 0α<1. Furthermore, the notions of Geraghty I and III contractions coincide when the metric space (X,ρ) is compact [28].

    Theorem 1. [26] Let (X,ρ) be a complete metric space and f:XX be a Geraghty contraction then there is a unique fixed point xX of f, and for each xX, limn+fn(x)=x.

    Theorem 2. Assume that wi:XX, for i=1,,n are Geraghty contractions of a RIFS, then there exists a unique set A=(A1,,An), called the attractor of the RIFS, such that W(A)=A and

    Ai=jI(i)wi(Aj),for i=1,,n.

    We identify A to the union of all Ai, that is, A=ni=1Ai.

    Proof. Let φ:(0,)(0,) be a nondecreasing function satisfying φ(t)t<1 for all t>0, and tφ(t)t is nonincreasing (or nondecreasing, or continuous), such that

    d(wi(x),wi(y))φ(d(x,y)),x,yX.

    Then, we have

    h(W(A),W(B))=h(ni=1wi(A),ni=1wi(B))max1inh(wi(A),wi(B))=max1in(max(d(wi(A),wi(B)),d(wi(B),wi(A))))max(φ(d(A,B)),φ(d(B,A)))φ(h(A,B)).

    Hence, the Hutchinson operator W is a Geraghty contraction and has a unique fixed compact set (Theorem 2.1). Using (2.3), this means that there exists n compact subsets Ai, for i=1,,n, such that A=(A1,,An) and

    Ai=jI(i)wi(Aj).

    An RIFS {X,w1,,wn,(pij)1i,jn} is said to satisfy the open set condition (OSC) if there exists bounded nonempty open sets U1,U2,,Un with the property that

    j,pji>0wi(Uj)Ui,wi(Uj)wi(Uk)=,for allkiljwithpjipkj>0. (2.5)

    When the RIFS satisies the OSC, we can compute the Hausdorff dimension of the attractor generated by the system. In fact, the attractor's Hausdorff dimension reflects how space-filling or complex the attractor is. Calculating the Hausdorff dimension of the attractor is a common method of quantifying its complexity. The condition (2.5) is satisfied in the case of RFIF in section three. It prevents the interpolated curve or fractal from collapsing into itself or having degenerate, overlapping segments. It ensures that the points used for interpolation are well separated and distinct, allowing for the generation of a meaningful and nondegenerate fractal or curve.

    Example 2. This example involves four transformations that take the square into itself. For 1i4, the range of wi is the quadrant of the square labeled i. If the transition probability pij is zero, no points on the attractor lie in the sub-quadrant labeled ij. While the structure of the attractor of the RIFS depends only on whether or not pij is nonzero and not on the values of the pijs, the coloring and texture of the image can be controlled through the probability values (see [16] for more details). To illustrate that, different choices for the transition probabilities are considered. Let us consider again the mappings wi,i=1,2,3,4 defined in Example 1. To construct the attractor, we apply one map of the IFS to a random selected point. Then, at each step, we apply a new map according to the probabilities of the transition of the matrix. In Figure 1 (a1, a2 and a3), we illustrate the attractors of RIFS equipped respectively with the following transition matrices :

    P1=(01000121201301313120012),P2=(14141414141414140131313001212),P3=(01201201201214141414120120).
    Figure 1.  The RIFSs equipped with the transition matrices P1, P2 and P3.

    Now, we consider ˜I={[0,1]2,˜wi,i=1,,4}, the RIFS equipped with the mappings

    ˜wi(x,y)=(aix+bi,cix+y1+βy+fi), for i=1,,4, (2.6)

    where β=0.01. These mappings are Geraghty contractions since they satisfy

    |cix1+y11+βy1cix2y21+βy2||ci||x1x2|+φ(|y1y2|), for all y1,y2[0,1],

    with φ(t)=t1+βt (this fact can be proved using Lemma 1). The attractors of ˜I, equipped with the transition matrices P1, P2 and P3, are illustrated in Figure 1 (b1, b2 and b3), respectively.

    For n2, let {(xi,yi)R2, i=0,,n} be a given data set, where x0<x1<<xn and yi[a,b] for all i=0,,n. For i=1,,n, we set I=[x0,xn] and Ii=[xi1,xi]. For each interval Ii, we define Di=[xl(i),xr(i)], with l(i),r(i){0,,n} such that l(i)<r(i) and xr(i)xl(i)>xixi1. For each i=1,,n, we consider Li:[xl(i),xr(i)][xi1,xi] as the homeomorphism given by Li(x)=aix+bi and Fi:Di×RR as continuous mappings satisfying

    Fi(xl(i),yl(i))=yi1andFi(xr(i),yr(i))=yi. (3.1)

    Assume that Fi is k-Lipschitz with respect to the first variable and a Geraghty contraction with respect to the second variable; that is

    |Fi(x,y)Fi(x,y)|k|xx|+φ(|yy|), for all x,xDi, and y,yR, (3.2)

    where k0 and φ:(0,)(0,) is a nondecreasing function such that, for any t>0, α(t):=φ(t)t<1 and the function α is nonincreasing (or nondecreasing, or continuous). We define wi:Di×RIi×R by wi(x,y)=(Li(x),Fi(x,y)). Set I(i)={j{1,,n} IjDi}.

    The proof of the following lemma mimics that of [9, Theorem 3.7] and [12, Theorem 3.3].

    Lemma 1. There exists a metric dθ in I×[a,b] equivalent to the Euclidean metric, such that for all i=1,,n, the mappings wi defined above, are Geraghty contractions with respect to dθ.

    Proof. We define a metric dθ on I×[a,b] by

    dθ((x,y),(x,y)):=|xx|+θ|yy|,

    where θ is a real number that is specified in the sequel. Then, for all (x,y),(x,y)I×[a,b], we have:

    dθ(wi(x,y),wi(x,y))=dθ((Li(x),Fi(x,y)),(Li(x),Fi(x,y)))=|Li(x)Li(x)|+θ|Fi(x,y)Fi(x,y)||ai||xx|+θk|xx|+θφ(|yy|)=(|ai|+θk)|xx|+θφ(|yy|)=(|ai|+θk+θφ(|yy|)|xx|+|yy|)|xx|+θφ(|yy|)|xx|+|yy||yy|(|ai|+θk+θφ(|xx|+|yy|)|xx|+|yy|)|xx|+θφ(|xx|+|yy|)|xx|+|yy||yy|(|ai|+θk+θ)|xx|+θφ(|xx|+|yy|)|xx|+|yy||yy|max(a+θk+θ,φ(|xx|+|yy|)|xx|+|yy|)dθ((x,y),(x,y)),

    with a=max1in|ai|. Set θ:=1a2(k+1), then 0<θ<1 and 0<a+θk+θ<1. Moreover, for t>0, let β(t):=max(a+θk+θ,φ(t)t). Then, β:(0,)[0,1) is a nonincreasing (or nondecreasing, or continuous), since tφ(t)t is a nonncreasing (or nondecreasing, or continuous). Furthermore, we obtain:

    dθ(wi(x,y),wi(x,y))β(d((x,y),(x,y)))dθ((x,y),(x,y)),

    where d((x,y),(x,y)):=|xx|+|yy|. Since 0<θ<1, then, for all (x,y)(x,y), we have:

    θ|xx|+θ|yy||xx|+θ|yy||xx|+|yy|.

    Thus,

    dθ((x,y),(x,y))d((x,y),(x,y))θ1dθ((x,y),(x,y)).

    If β:(0,)[0,1) is nonncreasing, then

    dθ(wi(x,y),wi(x,y))β(d((x,y),(x,y)))dθ((x,y),(x,y))β(dθ((x,y),(x,y)))dθ((x,y),(x,y)).

    If β:(0,)[0,1) is nondecreasing, then

    dθ(wi(x,y),wi(x,y))β(d((x,y),(x,y)))dθ((x,y),(x,y))β(θ1dθ((x,y),(x,y)))dθ((x,y),(x,y))β(dθ((x,y),(x,y)))dθ((x,y),(x,y)),

    where β:(0,)[0,1) is a nondecreasing function defined as β(t):=β(θ1t).

    If β:(0,)[0,1) is continuous, then

    dθ(wi(x,y),wi(x,y))β(d((x,y),(x,y)))dθ((x,y),(x,y))β(dθ((x,y),(x,y)))dθ((x,y),(x,y)),

    where β:(0,)[0,1) is a continuous function defined as β(t):=maxx[t,θ1t]β(x). Hence, wi are Geraghty contractions in (I×[a,b],dθ). Moreover, we have:

    min(1,θ)|xx|2+|yy|2min(1,θ)(|xx|+|yy|)dθ((x,y),(x,y))(1+θ)|xx|2+|yy|2.

    That is, dθ is equivalent to the Euclidean metric on I×[a,b]. In particular, (I×[a,b],dθ) is a complete metric space.

    Denote by C(I) the set of continuous functions f:I=[x0,xn][a,b]. Let C(I) be the set of continuous functions on I that pass through the interpolation points; that is

    C(I)={fC(I) | f(xi)=yi,  i=0,,n}.

    Then, C(I) and C(I) are complete metric spaces with respect to the uniform metric ρ defined by

    ρ(f1,f2)=maxxI|f1(x)f2(x)|.

    Define the Read-Bajraktarevic operator T on C(I) by

    Tf(x)=Fi(L1i(x),f(L1i(x))), for xIi and i=1,,n.

    It is easy to verify that Tf is continuous on I and that Tf(xi)=yi, for all i=0,,n.

    Our main result in this section is given by the following theorem. This extends [17, Corollaries 3.5] and [17, Corollaries 3.7] and the results given in [9,12] regarding the existence of FIF defined with Rakotch and Geraghty contractions.

    Theorem 3. For n2, let {(xi,yi); i=0,,n} be a given data set and let {Di×R,wi; i=1,,n} be the RIFS constructed above. Then, there exists a unique function ˜fC(I) such that ˜f(xi)=yi, for all i=0,,n. Moreover, let Gi be the graph of the restriction of ˜f to Ii, then G=(G1,,Gn) is the attractor of the RIFS {Di×R,wi; i=1,,n}, i.e.,

    Gi=jI(i)wi(Gj), for i=1,,n.

    Furthermore, the RFIF ˜f satisfies the functional equation

    ˜f(x)=Fi(L1i(x),˜fL1i(x)), for all xIi and i=1,,n.

    Proof. Let's first notice that the RIFS {Di×R,wi; i=1,,n} has a unique attractor by Lemma 1 and Theorem 2.1. Moreover, for f,gC(I), we have:

    ρ(Tf,Tg)=maxxI|Tf(x)Tg(x)|=max1inxIi|Tf(x)Tg(x)|=max1inxIi|Fi(L1i(x),f(L1i(x)))Fi(L1i(x),g(L1i(x)))|max1insupxIiφ(|f(L1i(x))g(L1i(x))|).

    Since, for each i{1,,n} and every xIi, we have:

    φ(|f(L1i(x))g(L1i(x))|)φ(maxxIi|f(L1i(x))g(L1i(x))|)φ(maxxI|f(x)g(x)|)=φ(ρ(f,g)),

    then, we obtain :

    ρ(Tf,Tg)φ(ρ(f,g)).

    Thus, T is a Geraghty contraction and has a unique fixed point ˜fC(I), i.e. there exists a unique continuous function ˜f that interpolates the points (xi,yi), i=0,,n and satisfies

    ˜f(x)=Fi(L1i(x),˜fL1i(x)), xIi, i=1,,n.

    Moreover, let G be the graph of f and let Gi be the graph of the restriction of f to Ii, for i=1,,n, then,

    G=ni=1Gi.

    Furthermore, we have:

    Gi={(x,˜f(x)); xIi}={(x,Fi(L1i(x),˜f(L1i(x)))) ;xIi}={(Li(x),Fi(x,˜f(x)));xDi}={wi(x,˜f(x));xDi}.

    Hence, (G1,,Gn) is the invariant set of W, i.e., the attractor of the RIFS {Di×R,wi; i=1,,n}, and we have:

    Gi=jI(i)wi(Gj), for i=1,,n.

    Example 3. Considering the interpolation data

    {(0,9);(0.25,11);(0.5,10);(0.75,11);(1,8)},

    we illustrate the RFIFs corresponding to ˜I equipped with the transition matrices P1, P2 and P3 given in Example 2 in Figure 2 (l1–l3), respectively. The subsequent figure (f1–f3) illustrates the RFIF equipped with affine transformations of Example 1 with the same respective transition matrices so that,

    Figure 2.  The RFIFs equipped with the transition matrices P1, P2 and P3.

    D1 D2 D3 D4
    P1 [12,1] [0,12] [14,34] [12,1]
    P2 [0,12] [0,34] [0,1] [0,1]
    P3 [12,1] [0,34] [12,1] [0,34].

    In this section, we will compare RFIF with the classical FIF technique in terms of error rates of the time-series representing the daily vaccinations data. We can show that the fractal interpolation method gives lower error rates than the cubic splines, which is often used to avoid the problem of Runge's phenomenon [29]. For this, we collect (from the website ourworldindata.org) daily vaccination data from the following four countries: Germany, Italy, Tunisia and France (see Table 1). As seen in Figure 3, the irregularities of this data is very visible and the reconstruction of the curve from the point of view of fractal interpolation is significantly relevant. This can be an effective way to recover missing information due to insufficient testing, and to analyze the complexity of the evolution of COVID-19 given the importance of vaccination data for each day. In addition, the variations of the FIF incorporate the behavior of the data (see Figure 4). Two methods of fractal interpolation namely linear interpolation, where

    Fi(x,y)=αy+cix+di,

    as well as fractal interpolation defined by Geraghty contractions, with

    Fi(x,y)=y1+βy+cix+di,

    are considered.

    Table 1.  To effectively test the goodness of fit of the interpolation techniques that we will use, the interpolation data for each country that we used is the sliced original data.
    Tun Fra Ita Ger
    total data obtained 273 669 671 670
    interpolation data 69 168 168 168

     | Show Table
    DownLoad: CSV
    Figure 3.  The data of new vaccinations for Tunisia, Italy, France and Germany.
    Figure 4.  The Graphs of linear FIFs and RFIFs induced from the interpolation data of Tunisia, Italy, France and Germany.

    A common way of measuring the quality of the fit of the interpolation is the normalized root mean square error (NRMSE), which is a statistical error indicator that facilitates the comparison between datasets with different scales and ensures a more universal measure of comparison. There exists various different methods of RMSE normalizations, and we will use

    NRMSE=ni=0(yi^yi)21nni=0yi,

    where yi is the ith observation of y, ˆyi is the predicted y value given by the interpolation technique and n is the number of total data considered.

    In Table 2, we present the results obtained for different parameters α and β concerning the daily vaccination data of France. The parameters α and β have been chosen arbitrarily. We can see that the RFIF is less sensitive to parameter change compared to the FIF which has big differences in the NMRSE for the various parameter values. This may be explained by the fact that other parameters are involved in the RFIF, namely, the coefficients of the stochastic matrix. Even though we did not get relevant information on the evolution of the disease, we can see the fractal structure of the epidemic curve. Moreover, the presence of parameters α and β help to get a wide variety of mappings for approximating problems. The obtained values show a slight improvement when the recurrent fractal interpolation is deployed. Nevertheless, there are still questions about optimal values of the entries of the stochastic matrix that can be managed to determine an optimal choice of the intervals Di (as defined in Section 3), as well as optimal choice of β (as given in [13,30] for the choice of the vertical scaling factor) so that the obtained curves fit as closely as possible to the real values.

    Table 2.  The NRMSE of fractal interpolation and recurrent fractal interpolation on France's data.
    Fractal interpolation Recurrent fractal interpolation
    Fi(x,y)=αy+cix+di α=0.1 273.819 264.536
    α=0.2 310.896 270.743
    α=0.3 379.131 281.771
    α=0.5 636.935 320.554
    Fi(x,y)=y1+βy+cix+di β=0.015 263.342 263.252
    β=0.025 263.283 263.283
    β=0.045 263.354 263.299
    β=0.1 263.278 263.345

     | Show Table
    DownLoad: CSV

    RIFS can be seen as improvements of the notion of IFS and were introduced in order to gain more flexibility in natural-shape generation and image compression. In this paper, we gave new conditions on the existence of the attractor of a new class of RIFS based on the condition of Geraghty contractions. Consequently, we constructed a new class of RFIF.

    The authors declare they have not used Artificial Intelligence (AI) tools in the creation of this article.

    The authors extend their appreciation to the Deanship of Scientific Research, Vice Presidency for Graduate Studies and Scientific Research at King Faisal University, Saudi Arabia, for financial support under the annual funding track [GRANT 4413].

    The authors declare no conflict of interest.



    [1] M. F. Barnsley, Fractal functions and interpolation, Constr. Approximation, 2 (1986), 303–329. https://doi.org/10.1007/BF01893434 doi: 10.1007/BF01893434
    [2] S. Dubuc, Interpolation through an iterative scheme, J. Math. Anal. appl., 114 (1986), 185–204. https://doi.org/10.1016/0022-247X(86)90077-6 doi: 10.1016/0022-247X(86)90077-6
    [3] S. Dubuc, Interpolation fractale, in Fractal Geometry and Analysis, (eds. J. Bélais and S. Dubuc), Kluwer Academic Publishers, Dordrecht, 1989.
    [4] R. Massopust, Fractal functions, fractal surfaces and wavelets, Acad. Press, 1994. https://doi.org/10.1016/B978-0-08-092458-8.50003-4
    [5] A. K. B. Chand, G. P. Kapoor, Generalized cubic spline fractal interpolation functions, SIAM J. Numer. Anal., 44 (2006), 655–676. https://doi.org/10.1137/040611070 doi: 10.1137/040611070
    [6] P. Viswanathan, A. K. B. Chand, M. A. Navascués, Fractal perturbation preserving fundamental shapes: Bounds on the scale factors, J. Math. Anal. Appl., 419 (2014), 804–817. https://doi.org/10.1016/j.jmaa.2014.05.019 doi: 10.1016/j.jmaa.2014.05.019
    [7] N. Vijender, Bernstein fractal trigonometric approximation, Acta Appl. Math., 159 (2018), 11–27. https://doi.org/10.1007/s10440-018-0182-1
    [8] N. Attia, M. Balegh, R. Amami, R. Amami, On the Fractal interpolation functions associated with Matkowski contractions, Electr. Res. Arch., 31 (2023), 4652–4668. https://doi.org/10.3934/era.2023238 doi: 10.3934/era.2023238
    [9] S. Ri, A new nonlinear fractal interpolation function, Fractals, 25 (2017). https://doi.org/10.1142/S0218348X17500633
    [10] S. Ri, New types of fractal interpolation surfaces, Chaos Solitons Fractal, 119 (2019), 291–297. https://doi.org/10.1016/j.chaos.2019.01.010 doi: 10.1016/j.chaos.2019.01.010
    [11] M. A. Navascués, C. Pacurar, V. Drakopoulos, Scale-free fractal interpolation, Fractal Fract., 6 (2022), 602. https://doi.org/10.3390/fractalfract6100602 doi: 10.3390/fractalfract6100602
    [12] J. Kim, H. Kim, H. Mun, Nonlinear fractal interpolation curves with function vertical scaling factors, Indian J. Pure Appl. Math., 51 (2020), 483–499. https://doi.org/10.1007/s13226-020-0412-x doi: 10.1007/s13226-020-0412-x
    [13] D. S. Mazel, M. H. Hayes, Using iterated function systems to model discrete sequences, IEEE Trans. Signal Process, 40 (1992) 1724–1734. https://doi.org/10.1109/78.143444
    [14] J. R. Price, Resampling and reconstruction with fractal interpolation functions, IEEE Signal Process. Lett., 5 (1998), 228–230. https://doi.org/10.1109/97.712106 doi: 10.1109/97.712106
    [15] R. F. Leonarduzzi, G. Schlotthauer, M. E. Torres, Wavelet leader based multifractal analysis of heart rate variability during myocardial ischaemia, in Annual International Conference of the IEEE Engineering in Medicine and Biology, (2010), 110–113. https://doi.org/10.1109/IEMBS.2010.5626091
    [16] M. F. Barnsley, Fractals Everywhere, Boston, MA: Academic Press, 1988.
    [17] M. F. Barnsley, J. H. Elton, D. P. Hardin, Recurrent iterated function systems, Constr. Approx., 5 (1989), 3–31. https://doi.org/10.1007/BF01889596 doi: 10.1007/BF01889596
    [18] M. F. Barnsley, A. E. Jacquin, Application of recurrent iterated function systems to images, Visual Commun. Image Process., 1001 (1988). https://doi.org/10.1117/12.968945
    [19] C. Yun, W. Metzler, M. Barski, Image compression predicated on recurrent iterated function systems, in 2nd International Conference on Mathematics and Statistics, (2008).
    [20] D. Luor, Fractal interpolation functions for random data sets, Chaos Solitons Fractals, 114 (2018), 256–263. https://doi.org/10.1016/j.chaos.2018.06.033 doi: 10.1016/j.chaos.2018.06.033
    [21] D. Luor, On the distributions of fractal functions that interpolate data points with Gaussian noise, Chaos Solitons Fractals, 135 (2020), 109743. https://doi.org/10.1016/j.chaos.2020.109743 doi: 10.1016/j.chaos.2020.109743
    [22] H. J. Ruan, J. C. Xiao, B. Yang, Existence and box dimension of general recurrent fractal fractal interpolation functions, Bull. Aust. Math. Soc., 103 (2021), 278–290. https://doi.org/10.1017/S0004972720001045 doi: 10.1017/S0004972720001045
    [23] N. Balasubramani, A. Gowrisankar, Affine recurrent fractal interpolation functions, Eur. Phys. J. Spec. Top, 230 (2021), 3765–3779. https://doi.org/10.1140/epjs/s11734-021-00306-7 doi: 10.1140/epjs/s11734-021-00306-7
    [24] M. F. Barnsley, S. Demko, Iterated function systems and the global construction of fractals, Proc. Roy. Soc. London Ser. A, 399 (1985), 243–275. https://doi.org/10.1098/rspa.1985.0057 doi: 10.1098/rspa.1985.0057
    [25] P. Bouboulis, Fractal Interpolation Surfaces: Theory and Applications in Image Compression, 2006.
    [26] M. A. Geraghty, On contractive mappings, Proc. Amer. Math. Soc., 40 (1973), 604–608. https://doi.org/10.1090/S0002-9939-1973-0334176-5
    [27] J. Jachymski, I. Jóźwik, Nonlinear contractive conditions: A comparison and related problems, Fixed Point Theory Appl., 77 (2007), 123–146. https://doi.org/10.4064/bc77-0-10 doi: 10.4064/bc77-0-10
    [28] F. Strobin, Attractors of generalized IFSs that are not attractors of IFSs, J. Math. Anal. Appl., 422 (2015), 99–108. https://doi.org/10.1016/j.jmaa.2014.08.029 doi: 10.1016/j.jmaa.2014.08.029
    [29] Najmuddin Ahmad, Khan Farah Deeba, Study of numerical accuracy in different spline interpolation techniques, Global J. Pure Appl. Math., 16 (2020), 687–693.
    [30] N. Vijender, V. Drakopoulos, On the Bernstein affine fractal interpolation curved lines and surfaces, Axioms, 9 (2020), 119. https://doi.org/10.3390/axioms9040119 doi: 10.3390/axioms9040119
  • This article has been cited by:

    1. Najmeddine Attia, Rim Amami, On linear transformation of generalized affine fractal interpolation function, 2024, 9, 2473-6988, 16848, 10.3934/math.2024817
    2. Najmeddine Attia, Hajer Jebali, On the box dimension of recurrent fractal interpolation functions defined with Matkowski contractions, 2024, 32, 0971-3611, 3453, 10.1007/s41478-024-00816-2
    3. Najmeddine Attia, Neji Saidi, Rim Amami, Rimah Amami, On the stability of Fractal interpolation functions with variable parameters, 2024, 9, 2473-6988, 2908, 10.3934/math.2024143
    4. Najmeddine Attia, Taoufik Moulahi, Rim Amami, Neji Saidi, Note on fractal interpolation function with variable parameters, 2023, 9, 2473-6988, 2584, 10.3934/math.2024127
    5. Alireza Khalili Golmankhaneh, Hamdullah Şevli, Carlo Cattani, Zoran Vidović, Fractal Hankel Transform, 2025, 9, 2504-3110, 135, 10.3390/fractalfract9030135
  • Reader Comments
  • © 2023 the Author(s), licensee AIMS Press. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0)
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Metrics

Article views(1404) PDF downloads(134) Cited by(5)

Figures and Tables

Figures(4)  /  Tables(2)

Other Articles By Authors

/

DownLoad:  Full-Size Img  PowerPoint
Return
Return

Catalog