Forecasting the gross domestic product using a weight direct determination neural network

Spyridon D. Mourtas; Emmanouil Drakonakis; Zacharias Bragoudakis; Spyridon D. Mourtas; Emmanouil Drakonakis; Zacharias Bragoudakis

doi:10.3934/math.20231237

AIMS Mathematics

2023, Volume 8, Issue 10: 24254-24273. doi: 10.3934/math.20231237

Previous Article Next Article

Research article Special Issues

Forecasting the gross domestic product using a weight direct determination neural network

1.
Department of Economics, Mathematics-Informatics and Statistics-Econometrics, National and Kapodistrian University of Athens, Sofokleous 1 Street, 10559 Athens, Greece
2.
Laboratory "Hybrid Methods of Modelling and Optimization in Complex Systems", Siberian Federal University, Prosp. Svobodny 79, 660041 Krasnoyarsk, Russia
3.
Bank of Greece, 10250 Athens, Greece
4.
National and Kapodistrian University of Athens, Greece

Received: 01 July 2023 Revised: 21 July 2023 Accepted: 28 July 2023 Published: 14 August 2023
MSC : 68T10, 65F20, 91B40

One of the most often used data science techniques in business, finance, supply chain management, production, and inventory planning is time-series forecasting. Due to the dearth of studies in the literature that propose unique weights and structure (WASD) based models for regression issues, the goal of this research is to examine the creation of such a model for time-series forecasting. Given that WASD neural networks have been shown to overcome limitations of traditional back-propagation neural networks, including slow training speed and local minima, a multi-function activated WASD for time-series (MWASDT) model that uses numerous activation functions, a new auto cross-validation method and a new prediction mechanism are proposed. The MWASDT model was used in forecasting the gross domestic product (GDP) for numerous nations to show off its exceptional capacity for learning and predicting. Compared to previous WASD-based models for time-series forecasting and traditional machine learning models that MATLAB has to offer, the new model has produced noticeably better forecasting results, especially on unseen data.

Keywords:

Citation: Spyridon D. Mourtas, Emmanouil Drakonakis, Zacharias Bragoudakis. Forecasting the gross domestic product using a weight direct determination neural network[J]. AIMS Mathematics, 2023, 8(10): 24254-24273. doi: 10.3934/math.20231237

Related Papers:

[1]	Osama Ala'yed, Belal Batiha, Diala Alghazo, Firas Ghanim . Cubic B-Spline method for the solution of the quadratic Riccati differential equation. AIMS Mathematics, 2023, 8(4): 9576-9584. doi: 10.3934/math.2023483
[2]	Azhar Iqbal, Abdullah M. Alsharif, Sahar Albosaily . Numerical study of non-linear waves for one-dimensional planar, cylindrical and spherical flow using B-spline finite element method. AIMS Mathematics, 2022, 7(8): 15417-15435. doi: 10.3934/math.2022844
[3]	Seherish Naz Khalid Ali Khan, Md Yushalify Misro . Hybrid B-spline collocation method with particle swarm optimization for solving linear differential problems. AIMS Mathematics, 2025, 10(3): 5399-5420. doi: 10.3934/math.2025249
[4]	Khalid K. Ali, K. R. Raslan, Amira Abd-Elall Ibrahim, Mohamed S. Mohamed . On study the fractional Caputo-Fabrizio integro differential equation including the fractional q-integral of the Riemann-Liouville type. AIMS Mathematics, 2023, 8(8): 18206-18222. doi: 10.3934/math.2023925
[5]	Yan Wu, Chun-Gang Zhu . Generating bicubic B-spline surfaces by a sixth order PDE. AIMS Mathematics, 2021, 6(2): 1677-1694. doi: 10.3934/math.2021099
[6]	Mohammad Tamsir, Neeraj Dhiman, Deependra Nigam, Anand Chauhan . Approximation of Caputo time-fractional diffusion equation using redefined cubic exponential B-spline collocation technique. AIMS Mathematics, 2021, 6(4): 3805-3820. doi: 10.3934/math.2021226
[7]	Rabia Noureen, Muhammad Nawaz Naeem, Dumitru Baleanu, Pshtiwan Othman Mohammed, Musawa Yahya Almusawa . Application of trigonometric B-spline functions for solving Caputo time fractional gas dynamics equation. AIMS Mathematics, 2023, 8(11): 25343-25370. doi: 10.3934/math.20231293
[8]	Osama Ala'yed, Ahmad Qazza, Rania Saadeh, Osama Alkhazaleh . A quintic B-spline technique for a system of Lane-Emden equations arising in theoretical physical applications. AIMS Mathematics, 2024, 9(2): 4665-4683. doi: 10.3934/math.2024225
[9]	Mostafa M. A. Khater, A. El-Sayed Ahmed . Strong Langmuir turbulence dynamics through the trigonometric quintic and exponential B-spline schemes. AIMS Mathematics, 2021, 6(6): 5896-5908. doi: 10.3934/math.2021349
[10]	Kai Wang, Guicang Zhang . Curve construction based on quartic Bernstein-like basis. AIMS Mathematics, 2020, 5(5): 5344-5363. doi: 10.3934/math.2020343

Abstract

1. Introduction

The study of nonlinear systems of singular initial value problems has recently attracted many mathematicians and physicists ^{[1,2,3,4,5,6,7,8,9,10,11,12]}. One of the systems in this category is the following Lane-Emden system of the form:

$\frac{{d}^{2}{\omega }_{1}\left(\tau \right)}{d{\tau }^{2}}+\frac{{\delta }_{1}}{\tau }\frac{d{\omega }_{1}\left(\tau \right)}{d\tau }+{{\mathrm{\hslash }}}_{1}\left({\omega }_{1}\right(\tau ),{\omega }_{2}(\tau \left)\right) = {{\mathrm{\aleph }}}_{1}\left(\tau \right),$

(1)

subject to

${\omega }_{1}\left(0\right) = {\varepsilon }_{1},{{\omega }^{{\mathrm{{\text{'}}}}}}_{1}\left(0\right) = 0,$

(2)

${\omega }_{2}\left(0\right) = {\vartheta }_{1},{{\omega }^{{\mathrm{{\text{'}}}}}}_{2}\left(0\right) = 0,$

where ${{\mathrm{\aleph }}}_{1},{{\mathrm{\aleph }}}_{2}$ are source functions, $\tau \in \left[{\mathrm{0,1}}\right],$ ${\delta }_{1},{\delta }_{2},{\varepsilon }_{1},$ and ${\vartheta }_{1}$ are real constants. The Lane-Emden equation, which was first investigated by astrophysicists Jonathan Homer Lane and Robert Emden, comes in various scientific applications in two kinds. For ${\mathrm{\hslash }}\left(\omega \left(\tau \right)\right) = {\omega }^{\gamma }\left(\tau \right)$ , (1) is called the Lane-Emden equation of index $\gamma$ , or of the first kind, that is a fundamental equation in the theory of star structure ^{[13,14,15,16,17,18,19]}. It represents the temperature variant of a spherical gas cloud subject to the principles of thermodynamics and its molecules' mutual attraction, see ^{[20,21,22,23,24,25,26]} and references therein. In astrophysics, the Lane-Emden equation represents Poisson's equation for the gravitational potential of a spherically symmetric, polytropic fluid, and self-gravitating at hydrostatic equilibrium ^[27]. Moreover, an important area of application for this type of equation is the study of species' diffusive transit and chemical reactions inside porous catalyst particles ^[27]. However, for ${\mathrm{\hslash }}\left(\omega \left(\tau \right)\right) = {e}^{\omega \left(\tau \right)}$ , (1) is called the Lane-Emden equation of the second kind that describes the dimensionless density distribution in a sphere of isothermal gas ^[28]. The singular behavior that arises at $\tau = 0$ is the primary difficulty of the Lane-Emden equations.

Recently, modeling a variety of physical and chemical phenomena, including chemical reactions, population evolution, and pattern formation leads to the system of Lane-Emden equations ^[7]. Therefore, numerous approaches have been proposed for the solutions of scalar and system of Lane-Emden equations ^{[1,2,10,11,12,12,13,14,15,17,18,42,43,44,45,46,47]}.

Spline methods employing piecewise polynomial functions have been demonstrated to be convenient methods for obtaining numerical solutions to many challenging models in science, engineering, and mathematics due to their simplicity of implementation and efficiency ^{[29,30,31,32,33,34]}. One of the well-known spline methods is the so-called B-spline (the "B" stands for basis) functions, which were first proposed by Schoenberg in 1946. The B-spline functions ^[35,36] have recently been a valuable tool in numerical computation, approximation theory, and image processing as they have various useful properties such as numerical stability of computations, local effects of coefficient changes, and built-in smoothness between neighboring polynomial pieces. The degrees of B-spline and the collocation points are the main factors that play a significant role in the execution of the technique and affect the outcomes to be achieved up to a required level of accuracy.

One of the most efficient and versatile techniques for obtaining approximate solutions is the cubic B-spline method (CBSM). The CBSM is a third-order piecewise polynomial constructed from a combination of recursive formulas referred to as the cubic B-spline basis. The derivation of the B-spline basis and the construction of the B-spline function are thoroughly discussed in ^[37,38]. In recent years, the CBSM has been successfully applied to various mathematical problems ^[39]. This demonstrates the effectiveness and usefulness of spline approaches through their numerous successful implementations. Therefore, this paper investigates the approximate solution of systems of the Lane-Emden equations using the CBSM.

This paper is organized as follows, in the next section, we present the basic preliminaries of the method. A short summary of cubic B-Spline method is presented in Section 3. We show the convergence analysis in Section 4. Finally, we present some numerical examples in Section 5.

2. Preliminaries

In this section, we introduce some basic facts regarding cubic B-spline approximation. Assume that the interval ${\mathrm{\Gamma }} = [\alpha ,\beta ]$ can be divided into $k$ subintervals [ ${\tau }_{i},{\tau }_{i+1}$ ] via

${\tau }_{i} = \alpha +i{\mathrm{\Lambda }},i = 0,{\mathrm{ }}\cdots ,k ,$

where

${\mathrm{\Lambda }} = ({\mathrm{\beta }}-{\mathrm{\alpha }})/k .$

The linear space of the cubic spline over the given partition is

${{{{\rm M}}}}_{3}\left(I\right) = \{\mu \left(\tau \right)\in {C}^{2}\left(I\right):{\mu \left(\tau \right)|}_{{I}_{i}}\in {P}_{3},i = 0,{\mathrm{ }}{\mathrm{ }}{\mathrm{ }}{\mathrm{ }}{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }},k-1\},$

where ${\mu \left(\tau \right)|}_{{I}_{i}}$ indicates the restriction of $\mu \left(\tau \right)$ to ${I}_{i}$ and ${P}_{3}$ indicates the set of cubic polynomials in one-variable. The dimension of linear space ${{{{\rm M}}}}_{3}\left(I\right)$ is $(k+3)$ . Extend ${\mathrm{\Gamma }} = [\alpha ,\beta ]$ to

$\stackrel{-}{{\mathrm{\Gamma }}} = [\alpha -3{\mathrm{\Lambda }},\beta +3{\mathrm{\Lambda }}]$

with the equidistant knots

${\tau }_{i} = \alpha +i{\mathrm{\Lambda }},i = -3,{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }},k+3 .$

The cubic B-spline function

${{{{\rm K}}}}_{i}\left(\tau \right),i = -1,{\mathrm{ }}{\mathrm{ }}{\mathrm{ }}{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }},k+1 ,$

is given by ^[40]

${{\rm K}_{{\text{i}}}}\left( \tau \right) = \left\{ {\begin{array}{*{20}{l}} {\frac{{{{(\tau - {\tau _i})}^3}}}{{6{\Lambda ^3}}}, \tau \in \left[ {{\tau _i},{\tau _{i + 1}}} \right],} \\ {\frac{{{{(\tau - {\tau _i})}^3}}}{{6{\Lambda ^3}}} - 2\frac{{{{\left( {\tau - {\tau _{i + 1}}} \right)}^3}}}{{3{\Lambda ^3}}},\; \tau \in \left[ {{\tau _{i + 1}},{\tau _{i + 2}}} \right],} \\ {\frac{{{{({\tau _{i + 4}} - \tau )}^3}}}{{6{\Lambda ^3}}} - 2\frac{{{{\left( {{\tau _{i + 3}} - \tau } \right)}^3}}}{{3{\Lambda ^3}}}, \tau \in \left[ {{\tau _{i + 2}},{\tau _{i + 3}}} \right],} \\ {\frac{{{{({\tau _{i + 4}} - \tau )}^3}}}{{6{\Lambda ^3}}}, \tau \in \left[ {{\tau _{i + 3}},{\tau _{i + 4}}} \right],} \\ {0, {{\text{else}}}.} \end{array}} \right.$

The ${{{{\rm K}}}}_{i}\left(\tau \right),i = -1,{\mathrm{ }}{\mathrm{ }}{\mathrm{ }}{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }},k+1$ , form basis splines of ${{{{\rm M}}}}_{3}\left(I\right)$ . The values of ${{{{\rm K}}}}_{i}\left(\tau \right),{{{{\rm K}}}}_{i}^{{\mathrm{{\text{'}}}}}\left(\tau \right)$ and ${{{{\rm K}}}}_{i}^{{\mathrm{{\text{'}}}}{\mathrm{{\text{'}}}}}\left(\tau \right)$ at the knots are recorded in Table 1.

Table 1. The values of

${{{{\rm K}}}}_{i}\left(\tau \right),{{{{\rm K}}}}_{i}^{{\text{'}}}\left(\tau \right)$ and

${{{{\rm K}}}}_{i}^{{\text{'}}{\text{'}}}\left(\tau \right)$ at the knots.

	${\tau }_{i-1}$	${\tau }_{i}$	${\tau }_{i+1}$	else
${{{{\rm K}}}}_{i}\left(\tau \right)$	$\frac{1}{6}$	$\frac{4}{6}$	$\frac{1}{6}$	0
${{{{\rm K}}}}_{i}^{{\text{'}}}\left(\tau \right)$	$\frac{1}{2{\mathrm{\Lambda }}}$	$0$	$-\frac{1}{2{\mathrm{\Lambda }}}$	0
${{{{\rm K}}}}_{i}^{{\text{'}}{\text{'}}}\left(\tau \right)$	$\frac{1}{{{\mathrm{\Lambda }}}^{2}}$	$\frac{-2}{{{\mathrm{\Lambda }}}^{2}}$	$\frac{1}{{{\mathrm{\Lambda }}}^{2}}$	0

| Show Table

DownLoad: CSV

For a sufficiently smooth function $\rho \left(\tau \right)$ , there is a unique cubic spline $\mu \left(\tau \right)\in {{{{\rm M}}}}_{3}\left(I\right)$ fulfilling the interpolation conditions

$\mu \left({\tau }_{i}\right) = \rho \left({\tau }_{i}\right),i = 0,{\mathrm{ }}{\mathrm{ }}{\mathrm{ }}{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }},k ,$

and

$\mu {\mathrm{{\text{'}}}}\left(\alpha \right) = \rho {\mathrm{{\text{'}}}}\left(\alpha \right),$

such that

$\mu \left(\tau \right) = \sum\limits _{i = -1}^{k+1}{{{\mathtt{λ}}}_{i}{{{\rm K}}}}_{i}\left(\tau \right),$

(3)

where ${{\mathtt{λ}}}_{i}{\mathrm{{\text{'}}}}$ s are constants to be estimated.

Using (3), we get

$\mu \left({\tau }_{j}\right) = \sum \limits_{i = -1}^{k+1}{{{\mathtt{λ}}}_{i}{{{\rm K}}}}_{i}\left({\tau }_{j}\right) = \frac{{{\mathtt{λ}}}_{j-1}+4{{\mathtt{λ}}}_{j}+{{\mathtt{λ}}}_{j+1}}{6},$

(4)

${\mu }^{{\mathrm{{\text{'}}}}}\left({\tau }_{j}\right) = \sum\limits _{i = -1}^{k+1}{{{\mathtt{λ}}}_{i}{{{\rm K}}}{\mathrm{{\text{'}}}}}_{i}\left({\tau }_{j}\right) = \frac{{{\mathtt{λ}}}_{j+1}-{{\mathtt{λ}}}_{j-1}}{2{\mathrm{\Lambda }}},$

(5)

${\mu }^{{\mathrm{{\text{'}}}}{\mathrm{{\text{'}}}}}\left({\tau }_{j}\right) = \sum\limits _{i = -1}^{k+1}{{{\mathtt{λ}}}_{i}{{{{\rm K}}}}^{{\mathrm{{\text{'}}}}{\mathrm{{\text{'}}}}}}_{i}\left({\tau }_{j}\right) = \frac{{{\mathtt{λ}}}_{j-1}-2{{\mathtt{λ}}}_{j}+{{\mathtt{λ}}}_{j+1}}{{{\mathrm{\Lambda }}}^{2}}.$

(6)

Equations (4)–(6) are the most important relations in deriving the CBSM.

3. Cubic B-spline method

In this section, we present the cubic B-spline method for (1) and (2). Let

${\mu }_{1}\left(\tau \right) = \sum\limits _{i = -1}^{k+1}{{{\mathtt{λ}}}_{i}{{{\rm K}}}}_{i}\left(\tau \right)$

and

${\mu }_{2}\left(\tau \right) = \sum\limits _{i = -1}^{k+1}{{{\mathtt{η}}}_{i}{{{\rm K}}}}_{i}\left(\tau \right)$

denote the approximate solutions of system (1) and (2) of ${\omega }_{1}\left(\tau \right)$ and ${\omega }_{2}\left(\tau \right)$ , respectively. To overcome the singularity of (1) at $\tau = 0$ , we apply the L'Hôpital's rule to the second term as $\tau$ approaches zero, to achieve

$\left\{\begin{array}{c}(1+{\delta }_{1})\frac{{d}^{2}{\omega }_{1}\left(\tau \right)}{d{\tau }^{2}}+{{\mathrm{\hslash }}}_{1}\left({\omega }_{1}\right(\tau ),{\omega }_{2}(\tau \left)\right) = {{\mathrm{\aleph }}}_{1}\left(\tau \right),\\ (1+{\delta }_{2})\frac{{d}^{2}{\omega }_{2}\left(\tau \right)}{d{\tau }^{2}}+{{\mathrm{\hslash }}}_{2}\left({\omega }_{1}\right(\tau ),{\omega }_{2}(\tau \left)\right) = {{\mathrm{\aleph }}}_{2}\left(\tau \right),\end{array}\right.{\mathrm{f}}{\mathrm{o}}{\mathrm{r}}\;\tau = 0,\\\left\{\begin{array}{c}\frac{{d}^{2}{\omega }_{1}\left(\tau \right)}{d{\tau }^{2}}+\frac{{\delta }_{1}}{\tau }\frac{d{\omega }_{1}\left(\tau \right)}{d\tau }+{{\mathrm{\hslash }}}_{1}\left({\omega }_{1}\right(\tau ),{\omega }_{2}(\tau \left)\right) = {{\mathrm{\aleph }}}_{1}\left(\tau \right),\\ \frac{{d}^{2}{\omega }_{2}\left(\tau \right)}{d{\tau }^{2}}+\frac{{\delta }_{2}}{\tau }\frac{d{\omega }_{2}\left(\tau \right)}{d\tau }+{{\mathrm{\hslash }}}_{2}\left({\omega }_{1}\right(\tau ),{\omega }_{2}(\tau \left)\right) = {{\mathrm{\aleph }}}_{2}\left(\tau \right),\end{array}\right.{\mathrm{f}}{\mathrm{o}}{\mathrm{r}}\;\tau \ne 0.$

(7)

By discretizing (7), we get

$\left\{\begin{array}{c}(1+{\delta }_{1})\frac{{d}^{2}{\omega }_{1}\left({\tau }_{0}\right)}{d{\tau }^{2}}+{{\mathrm{\hslash }}}_{1}\left({\omega }_{1}\right({\tau }_{0}),{\omega }_{2}({\tau }_{0}\left)\right) = {{\mathrm{\aleph }}}_{1}\left({\tau }_{0}\right),\\ (1+{\delta }_{2})\frac{{d}^{2}{\omega }_{2}\left({\tau }_{0}\right)}{d{\tau }^{2}}+{{\mathrm{\hslash }}}_{2}\left({\omega }_{1}\right({\tau }_{0}),{\omega }_{2}({\tau }_{0}\left)\right) = {{\mathrm{\aleph }}}_{2}\left({\tau }_{0}\right),\end{array}\right.\\\left\{\begin{array}{c}\frac{{d}^{2}{\omega }_{1}\left({\tau }_{j}\right)}{d{\tau }^{2}}+\frac{{\delta }_{1}}{{\tau }_{j}}\frac{d{\omega }_{1}\left({\tau }_{j}\right)}{d\tau }+{{\mathrm{\hslash }}}_{1}\left({\omega }_{1}\right({\tau }_{j}),{\omega }_{2}({\tau }_{j}\left)\right) = {{\mathrm{\aleph }}}_{1}\left({\tau }_{j}\right),\\ \frac{{d}^{2}{\omega }_{2}\left({\tau }_{j}\right)}{d{\tau }^{2}}+\frac{{\delta }_{2}}{{\tau }_{j}}\frac{d{\omega }_{2}\left({\tau }_{j}\right)}{d\tau }+{{\mathrm{\hslash }}}_{2}\left({\omega }_{1}\right({\tau }_{j}),{\omega }_{2}({\tau }_{j}\left)\right) = {{\mathrm{\aleph }}}_{2}\left({\tau }_{j}\right),\end{array}\right.$

(8)

where $j = 1,\cdots ,k$ . By using (2)–(5), (8) becomes

$\left\{\begin{array}{c}\left(1+{\delta }_{1}\right)\left(\frac{{{\mathtt{λ}}}_{-1}-2{{\mathtt{λ}}}_{0}+{{\mathtt{λ}}}_{1}}{{{\mathrm{\Lambda }}}^{2}}\right)+{{\mathrm{\hslash }}}_{1}({\varepsilon }_{1},{\vartheta }_{1}) = {{\mathrm{\aleph }}}_{1}\left({\tau }_{0}\right),\\ \left(1+{\delta }_{2}\right)\left(\frac{{{\mathtt{η}}}_{-1}-2{{\mathtt{η}}}_{0}+{{\mathtt{η}}}_{1}}{{{\mathrm{\Lambda }}}^{2}}\right)+{{\mathrm{\hslash }}}_{2}({\varepsilon }_{1},{\vartheta }_{1}) = {{\mathrm{\aleph }}}_{2}\left({\tau }_{0}\right),\end{array}\right.$

$\left\{\begin{array}{c}\begin{array}{c}\left(\frac{{\lambda }_{i-1}-2{\lambda }_{i}+{\lambda }_{i+1}}{{\Lambda }^{2}}\right)+\frac{{\delta }_{1}}{{\tau }_{j}}\left(\frac{{\lambda }_{j+1}-{\lambda }_{j-1}}{2\Lambda }\right)+{\hslash }_{1}\left(\frac{{\lambda }_{i-1}+4{\lambda }_{i}+{\lambda }_{i+1}}{6}{\text{,}}\frac{{\eta }_{i-1}+4{\eta }_{i}+{\eta }_{i+1}}{6}\right)\\ = {\aleph }_{1}\left({\tau }_{j}\right),\hspace{0.33em}j = 1,\cdots ,k,\end{array}\\ \begin{array}{c}\left(\frac{{\eta }_{i-1}-2{\eta }_{i}+{\eta }_{i+1}}{{\Lambda }^{2}}\right)+\frac{{\delta }_{2}}{{\tau }_{j}}\left(\frac{{\eta }_{j+1}-{\eta }_{j-1}}{2\Lambda }\right)+{\hslash }_{2}\left(\frac{{\lambda }_{i-1}+4{\lambda }_{i}+{\lambda }_{i+1}}{6}{\text{,}}\frac{{\eta }_{i-1}+4{\eta }_{i}+{\eta }_{i+1}}{6}\right)\\ = {\aleph }_{2}\left({\tau }_{j}\right),\hspace{0.33em}\hspace{0.33em}j = 1,\cdots ,k.\end{array}\end{array}\right.$

(9)

The initial conditions (2) as well provide the following four equations

${\omega }_{1}\left(0\right) = {\varepsilon }_{1} = \frac{{{\mathtt{λ}}}_{-1}+4{{\mathtt{λ}}}_{0}+{{\mathtt{λ}}}_{1}}{6},$

(10)

${\omega {\text{'}}}_{1}\left(0\right) = 0 = \frac{{{\mathtt{λ}}}_{1}-{{\mathtt{λ}}}_{-1}}{2{\mathrm{\Lambda }}},$

(11)

${\omega }_{2}\left(0\right) = {\vartheta }_{1} = \frac{{{\mathtt{η}}}_{-1}+4{{\mathtt{η}}}_{0}+{{\mathtt{η}}}_{1}}{6},$

(12)

${\omega {\text{'}}{\text{'}}}_{2}\left(0\right) = 0 = \frac{{{\mathtt{η}}}_{1}-{{\mathtt{η}}}_{-1}}{2{\mathrm{\Lambda }}}.$

(13)

Equations (9)–(13) give us $2(k+3)$ nonlinear equations with ${{\mathtt{λ}}}_{i}$ and ${{\mathtt{η}}}_{i},$ $i = -1,\dots ,k+1$ as unknowns. Upon solving this system, we determine the coefficients of

${\mu }_{1}\left(\tau \right) = \sum\limits _{i = -1}^{k+1}{{{\mathtt{λ}}}_{i}{{{\rm K}}}}_{i}\left(\tau \right)$

and

${\mu }_{2}\left(\tau \right) = \sum\limits _{i = -1}^{k+1}{{{\mathtt{η}}}_{i}{{{\rm K}}}}_{i}\left(\tau \right).$

4. Convergence analysis

In this section, we analyze the convergence for the proposed method. For this purpose, we assume ${\omega }_{1}\left(\tau \right),{\omega }_{2}\left(\tau \right)\in {C}^{5}\left[{\mathrm{0,1}}\right].$ From (3)–(5), we get ^[39]

${\mathrm{\Lambda }}\left[{\mu }_{j}^{{\mathrm{{\text{'}}}}}\left({\tau }_{i-1}\right)+4{\mu }_{j}^{{\mathrm{{\text{'}}}}}\left({\tau }_{i}\right)+{\mu }_{j}^{{\mathrm{{\text{'}}}}}\left({\tau }_{i+1}\right)\right] = 3\left[{\omega }_{j}\left({\tau }_{i+1}\right)-{\omega }_{j}\left({\tau }_{i-1}\right)\right],$

(14)

${{\mathrm{\Lambda }}}^{2}{\mu }_{j}^{{\mathrm{{\text{'}}}}{\mathrm{{\text{'}}}}}\left({\tau }_{i}\right) = 6\left[{\mu }_{j}\left({\tau }_{i+1}\right)-{\mu }_{j}\left({\tau }_{i}\right)\right]-2{\mathrm{\Lambda }}\left[2{\mu }_{j}^{{\mathrm{{\text{'}}}}}\left({\tau }_{i}\right)+{\mu }_{j}^{{\mathrm{{\text{'}}}}}\left({\tau }_{i+1}\right)\right],$

(15)

$j = {\mathrm{1,2}}.$ Using the shifting operator, $E\left({\mu }_{j}\left({\tau }_{i}\right)\right) = {\mu }_{j}\left({\tau }_{i+1}\right),$ (14) may be written as

$\frac{{\mathrm{\Lambda }}}{6}\left({E}^{-1}+4+E\right){\mu }_{j}^{{\mathrm{{\text{'}}}}}\left({\tau }_{i}\right) = \frac{1}{2}\left(E-{E}^{-1}\right){\omega }_{j}\left({\tau }_{i}\right),$

(16)

$j = {\mathrm{1,2}}.$ Since $E = {e}^{{\mathrm{\Lambda }}D}$ and $D\equiv d/d\tau$ , one can get

${e}^{{\mathrm{\Lambda }}D}+{e}^{-{\mathrm{\Lambda }}D} = 2\sum\limits _{k = 0}^{{\mathrm{\infty }}}\frac{{\left({\mathrm{\Lambda }}D\right)}^{2k}}{\left(2k\right)!},\\{e}^{{\mathrm{\Lambda }}D}-{e}^{-{\mathrm{\Lambda }}D} = 2\sum\limits _{k = 0}^{{\mathrm{\infty }}}\frac{{\left({\mathrm{\Lambda }}D\right)}^{2k+1}}{\left(2k+1\right)!}.$

(17)

Therefore, using (17), (16) can be expressed as

$\left(1+\frac{1}{3}\sum\limits _{k = 1}^{{\mathrm{\infty }}}\frac{{\left({\mathrm{\Lambda }}D\right)}^{2k}}{\left(2k\right)!}\right){\mu }_{j}^{{\mathrm{{\text{'}}}}}\left({\tau }_{i}\right) = \left(\sum\limits _{k = 0}^{{\mathrm{\infty }}}\frac{{\left({\mathrm{\Lambda }}D\right)}^{2k+1}}{(2k+1)!}\right){\omega }_{j}\left({\tau }_{i}\right).$

(18)

Simplifying (18) gives

${\mu }_{j}^{{\text{'}}}\left({\tau }_{i}\right) = \left(\sum\limits _{k = 0}^{\infty }\frac{{\left(\Lambda D\right)}^{2k+1}}{(2k+1)!}\right){\left(1+\frac{1}{3}\sum\limits _{k = 1}^{\infty }\frac{{\left(\Lambda D\right)}^{2k}}{\left(2k\right)!}\right)}^{-1}{\omega }_{j}\left({\tau }_{i}\right)\\{\text{ = (D+}}\frac{{{\mathrm{\Lambda }}}^{2}{{\mathrm{D}}}^{3}}{3!}+\frac{{{\mathrm{\Lambda }}}^{4}{{\mathrm{D}}}^{5}}{5!}+\cdots )\left(1-\left(\frac{{{\mathrm{\Lambda }}}^{2}{{\mathrm{D}}}^{2}}{6}+\frac{{{\mathrm{\Lambda }}}^{4}{{\mathrm{D}}}^{4}}{72}+\cdots \right)\right.\\\left.+{\left(\frac{{\Lambda }^{2}{D}^{2}}{6}+\frac{{\Lambda }^{4}{D}^{4}}{72}+\cdots \right)}^{2}+\cdots \right){\omega }_{j}\left({\tau }_{i}\right)\\ = D\left(1-\frac{{\Lambda }^{4}{D}^{4}}{180}+\frac{{\Lambda }^{6}{D}^{6}}{1512}-\cdots \right){\omega }_{j}\left({\tau }_{i}\right).$

(19)

Therefore,

${\mu }_{j}^{{\mathrm{{\text{'}}}}}\left({\tau }_{i}\right) = {\omega }_{j}^{{\mathrm{{\text{'}}}}}\left({\tau }_{i}\right)-\frac{{{\mathrm{\Lambda }}}^{4}}{180}{\omega }_{j}^{\left(5\right)}\left({\tau }_{i}\right)+\cdots .$

(20)

Similarly, (15) gives

${\mu }_{j}^{{\mathrm{{\text{'}}}}{\mathrm{{\text{'}}}}}\left({\tau }_{i}\right) = {\omega }_{j}^{{\text{'}}{\text{'}}}\left({\tau }_{i}\right)-\frac{1}{12}{{\mathrm{\Lambda }}}^{2}{\omega }_{j}^{\left(4\right)}\left({\tau }_{i}\right)+\frac{1}{360}{{\mathrm{\Lambda }}}^{4}{\omega }_{j}^{\left(6\right)}\left({\tau }_{i}\right)+O\left({{\mathrm{\Lambda }}}^{6}\right),$

(21)

At this point, the error functions ${e}_{j}\left(\tau \right),j = {\mathrm{1,2}},$ are stated as follows:

${e}_{1}\left({\tau }_{i}\right) = {\aleph }_{1}\left({\tau }_{j}\right)-\frac{{d}^{2}{\omega }_{1}\left({\tau }_{j}\right)}{d{\tau }^{2}}-\frac{{\delta }_{1}}{{\tau }_{j}}\frac{d{\omega }_{1}\left({\tau }_{j}\right)}{d\tau }-{\hslash }_{1}\left({\omega }_{1}\right({\tau }_{j}),{\omega }_{2}({\tau }_{j}\left)\right)\\ = \frac{{d}^{2}{\mu }_{1}\left({\tau }_{j}\right)}{d{\tau }^{2}}+\frac{{\delta }_{1}}{{\tau }_{j}}\frac{d{\mu }_{1}\left({\tau }_{j}\right)}{d\tau }+{\hslash }_{1}\left({\mu }_{1}\left({\tau }_{j}\right),{\mu }_{2}\left({\tau }_{j}\right)\right)\\ -\frac{{d}^{2}{\omega }_{1}\left({\tau }_{j}\right)}{d{\tau }^{2}}-\frac{{\delta }_{1}}{{\tau }_{j}}\frac{d{\omega }_{1}\left({\tau }_{j}\right)}{d\tau }-{\hslash }_{1}\left({\omega }_{1}\right({\tau }_{j}),{\omega }_{2}({\tau }_{j}\left)\right)\\ = \left[\frac{{d}^{2}{\mu }_{1}\left({\tau }_{j}\right)}{d{\tau }^{2}}-\frac{{d}^{2}{\omega }_{1}\left({\tau }_{j}\right)}{d{\tau }^{2}}\right]+\frac{{\delta }_{1}}{{\tau }_{j}}\left[\frac{d{\mu }_{1}\left({\tau }_{j}\right)}{d\tau }-\frac{d{\omega }_{1}\left({\tau }_{j}\right)}{d\tau }\right],\\{e}_{2}\left({\tau }_{i}\right) = {\aleph }_{2}\left({\tau }_{j}\right)-\frac{{d}^{2}{\omega }_{2}\left({\tau }_{j}\right)}{d{\tau }^{2}}-\frac{{\delta }_{2}}{{\tau }_{j}}\frac{d{\omega }_{2}\left({\tau }_{j}\right)}{d\tau }-{\hslash }_{2}\left({\omega }_{1}\right({\tau }_{j}),{\omega }_{2}({\tau }_{j}\left)\right)\\ = \frac{{d}^{2}{\mu }_{2}\left({\tau }_{j}\right)}{d{\tau }^{2}}+\frac{{\delta }_{2}}{{\tau }_{j}}\frac{d{\mu }_{2}\left({\tau }_{j}\right)}{d\tau }+{\hslash }_{2}\left({\mu }_{1}\left({\tau }_{j}\right),{\mu }_{2}\left({\tau }_{j}\right)\right)\\-\frac{{d}^{2}{\omega }_{2}\left({\tau }_{j}\right)}{d{\tau }^{2}}-\frac{{\delta }_{2}}{{\tau }_{j}}\frac{d{\omega }_{2}\left({\tau }_{j}\right)}{d\tau }-{\hslash }_{2}\left({\omega }_{1}\right({\tau }_{j}),{\omega }_{2}({\tau }_{j}\left)\right)\\ = \left[\frac{{d}^{2}{\mu }_{2}\left({\tau }_{j}\right)}{d{\tau }^{2}}-\frac{{d}^{2}{\omega }_{2}\left({\tau }_{j}\right)}{d{\tau }^{2}}\right]+\frac{{\delta }_{2}}{{\tau }_{j}}\left[\frac{d{\mu }_{2}\left({\tau }_{j}\right)}{d\tau }-\frac{d{\omega }_{2}\left({\tau }_{j}\right)}{d\tau }\right],$

(22)

where $j = 1,{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }}.{\mathrm{ }},k$ . Substitute (20) and (21) in (22) yields

${‖{e}_{1}\left({\tau }_{i}\right)‖}_{{\mathrm{\infty }}} = O\left({{\mathrm{\Lambda }}}^{2}\right),$

${‖{e}_{2}\left({\tau }_{i}\right)‖}_{{\mathrm{\infty }}} = O\left({{\mathrm{\Lambda }}}^{2}\right).$

(23)

For i = 0, we get

${e}_{1}\left({\tau }_{0}\right) = {\aleph }_{1}\left({\tau }_{0}\right)-\left(1+{\delta }_{1}\right)\frac{{d}^{2}{\omega }_{1}\left({\tau }_{0}\right)}{d{\tau }^{2}}-{\hslash }_{1}\left({\omega }_{1}\left({\tau }_{0}\right),{\omega }_{2}\left({\tau }_{0}\right)\right)\\ = \left(1+{\delta }_{1}\right)\frac{{d}^{2}{\mu }_{1}\left({\tau }_{0}\right)}{d{\tau }^{2}}+{\hslash }_{1}\left({\mu }_{1}\left({\tau }_{0}\right),{\mu }_{2}\left({\tau }_{0}\right)\right)\\-\left(1+{\delta }_{1}\right)\frac{{d}^{2}{\omega }_{1}\left({\tau }_{0}\right)}{d{\tau }^{2}}-{\hslash }_{1}({\omega }_{1}\left({\tau }_{0}\right),{\omega }_{2}\left({\tau }_{0}\right)\\ = \left(1+{\delta }_{1}\right)\left[\frac{{d}^{2}{\mu }_{1}\left({\tau }_{0}\right)}{d{\tau }^{2}}-\frac{{d}^{2}{\omega }_{1}\left({\tau }_{0}\right)}{d{\tau }^{2}}\right],\\{e}_{2}\left({\tau }_{0}\right) = {\aleph }_{2}\left({\tau }_{0}\right)-(1+{\delta }_{2})\frac{{d}^{2}{\omega }_{2}\left({\tau }_{0}\right)}{d{\tau }^{2}}-{\hslash }_{2}({\omega }_{1}\left({\tau }_{0}\right),{\omega }_{2}\left({\tau }_{0}\right)) \\ = \left(1+{\delta }_{2}\right)\frac{{d}^{2}{\omega }_{2}\left({\tau }_{0}\right)}{d{\tau }^{2}}+{\hslash }_{2}\left({\mu }_{1}\left({\tau }_{0}\right),{\mu }_{2}\left({\tau }_{0}\right)\right)\\ -\left(1+{\delta }_{2}\right)\frac{{d}^{2}{\omega }_{2}\left({\tau }_{0}\right)}{d{\tau }^{2}}-{\hslash }_{1}({\omega }_{1}\left({\tau }_{0}\right),{\omega }_{2}\left({\tau }_{0}\right)\\ = (1+{\delta }_{2})\left[\frac{{d}^{2}{\omega }_{2}\left({\tau }_{0}\right)}{d{\tau }^{2}}-\frac{{d}^{2}{\omega }_{2}\left({\tau }_{0}\right)}{d{\tau }^{2}}\right].$

(24)

Using (21) in (24), we have

${‖{e}_{1}\left({\tau }_{0}\right)‖}_{{\mathrm{\infty }}} = O\left({{\mathrm{\Lambda }}}^{2}\right),\\{‖{e}_{2}\left({\tau }_{0}\right)‖}_{{\mathrm{\infty }}} = O\left({{\mathrm{\Lambda }}}^{2}\right).$

(25)

Therefore, from (23) and (25), the truncation error for the considered system is $O\left({{\mathrm{\Lambda }}}^{2}\right)$ .

5. Numerical results

In this section, we present the numerical solution to (1) and (2) using the cubic B-spline technique. Several problems are examined to prove the accuracy and efficiency of the proposed method using the absolute errors between the approximate solutions and the exact solutions ( $\left|{e}_{j}\left(\tau \right)\right|)$ for various k. All the results are calculated by using MATHEMATICA 12.

Problem 1. Consider the following system ^[24]

$\frac{{d}^{2}{\omega }_{1}\left(\tau \right)}{d{\tau }^{2}}+\frac{3}{\tau }\frac{d{\omega }_{1}\left(\tau \right)}{d\tau }-4({\omega }_{1}\left(\tau \right)+{\omega }_{2}\left(\tau \right)) = 0,\\\frac{{d}^{2}{\omega }_{2}\left(\tau \right)}{d{\tau }^{2}}+\frac{2}{\tau }\frac{d{\omega }_{2}\left(\tau \right)}{d\tau }+3({\omega }_{1}\left(\tau \right)+{\omega }_{2}\left(\tau \right)) = 0,$

(26)

subject to

${\omega }_{1}\left(0\right) = 1,{\omega }_{1}^{{\text{'}}}\left(0\right) = 0,\\{\omega }_{2}\left(0\right) = 1,{\omega }_{2}^{{\text{'}}}\left(0\right) = 0,$

(27)

where the exact solutions are ${\omega }_{1}\left(\tau \right) = 1+{\tau }^{2}$ and ${\omega }_{2}\left(\tau \right) = 1-{\tau }^{2}$ . In this example, we use ${\mathrm{\Lambda }} = 0.1$ . The numerical results and the exact solution at the grid points are listed in Table 2. We can conclude from Table 2 that the obtained numerical results are in excellent agreement with the exact solutions. We note that for this problem, our results are exact and the achieved errors are only due to round-off calculations. The CPU time for this problem, with ${\mathrm{\Lambda }} = 0.1$ , is 0.0156s.

Table 2. Numerical results for Problem 1.

$\tau$	${\omega }_{1}\left(\tau \right)$	${\mu }_{1}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.1$ )	$\left\|{e}_{1}\left(\tau \right)\right\|$	${\omega }_{2}\left(\tau \right)$	${\mu }_{2}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.1$ )	$\left\|{e}_{2}\left(\tau \right)\right\|$
$0.0$	1	$1$	$0$	$1$	$1$	$2.22045\times {10}^{-16}$
$0.1$	1.01	$1.01$	$0$	0.99	0.99	$2.22045\times {10}^{-16}$
$0.2$	1.04	$1.04$	$0$	0.96	0.96	$2.22045\times {10}^{-16}$
$0.3$	1.09	$1.09$	$2.22045\times {10}^{-16}$	0.91	0.91	$2.22045\times {10}^{-16}$
$0.4$	1.16	1.16	$2.22045\times {10}^{-16}$	0.84	0.84	$3.33067\times {10}^{-16}$
$0.5$	1.25	1.25	$2.22045\times {10}^{-16}$	0.75	0.75	$3.33067\times {10}^{-16}$
$0.6$	1.36	1.36	$4.44089\times {10}^{-16}$	0.64	0.64	$3.33067\times {10}^{-16}$
$0.7$	1.49	1.49	$2.22045\times {10}^{-16}$	0.51	0.51	$4.44089\times {10}^{-16}$
$0.8$	1.61	1.61	$0$	0.36	0.36	$4.44089\times {10}^{-16}$
$0.9$	1.81	1.81	$2.22045\times {10}^{-16}$	0.19	0.19	$5.55112\times {10}^{-16}$
$1.0$	2	2	$0$	0	5.64363 $\times {10}^{-16}$	$5.64363\times {10}^{-16}$

| Show Table

DownLoad: CSV

Problem 2. Consider the following system ^[27]

${\omega }_{1}^{{\text{'}}{\text{'}}}\left(\tau \right)+\frac{2}{\tau }{\omega }_{1}^{{\text{'}}}\left(\tau \right)-\left(4{\tau }^{2}+6\right){\omega }_{1}\left(\tau \right)+{\omega }_{2}\left(\tau \right) = {\tau }^{4}-{\tau }^{3},\\{\omega }_{2}^{{\text{'}}{\text{'}}}\left(\tau \right)+\frac{8}{\tau }{\omega }_{2}^{{\text{'}}}\left(\tau \right)+{\omega }_{1}\left(\tau \right)+\tau {\omega }_{2}\left(\tau \right) = {e}^{{\tau }^{2}}+{\tau }^{5}-{\tau }^{4}+44{\tau }^{2}-30\tau ,$

(28)

subject to

${\omega }_{1}\left(0\right) = 1,{\omega }_{1}^{{\text{'}}}\left(0\right) = 0,\\{\omega }_{2}\left(0\right) = 0,{\omega }_{2}^{{\text{'}}}\left(0\right) = 0,$

(29)

where the exact solutions are

${\omega }_{1}\left(\tau \right) = {e}^{{\tau }^{2}}$

and

${\omega }_{2}\left(\tau \right) = {\tau }^{4}-{\tau }^{3}.$

The obtained numerical and exact solutions, with different values of ${\mathrm{\Lambda }}$ , are depicted in Figure 1. The absolute errors between exact and numerical results, with ${\mathrm{\Lambda }} = 0.1$ and $0.01$ , are presented in Tables 3 and 4. In Table 5, we compare the maximum absolute error of CBSM with those of ^[27]. Our results seem to be better than those of ^[27]. The CPU time for this problem, with ${\mathrm{\Lambda }} = 0.1$ and $0.01$ , are 0.0156s and 0.0625s respectively.

Figure 1. Absolute error functions for Problem 2.

DownLoad: Full-Size Img PowerPoint

Table 3. Numerical result for approximate solution of

${\omega }_{1}\left(\tau \right)$ in Problem 2.

$\tau$	${\omega }_{1}\left(\tau \right)$	${\mu }_{1}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.1$ )	$\left\|{e}_{1}\left(\tau \right)\right\|$	${\mu }_{1}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.01$ )	$\left\|{e}_{1}\left(\tau \right)\right\|$
$0.0$	$1$	$1$	$0$	$1$	$0$
$0.1$	$1.01005$	$1.01008$	$3.38838\times {10}^{-5}$	$1.01005$	$1.71898\times {10}^{-7}$
$0.2$	$1.04081$	$1.04091$	$9.51843\times {10}^{-5}$	$1.04081$	$7.1747\times {10}^{-7}$
$0.3$	1.09417	$1.09438$	$2.01533\times {10}^{-4}$	$1.09417$	$1.75608\times {10}^{-6}$
$0.4$	1.17351	1.17389	$3.80997\times {10}^{-4}$	$1.17351$	$3.5098\times {10}^{-6}$
$0.5$	1.28403	1.2847	$6.71612\times {10}^{-4}$	$1.28403$	$6.3527\times {10}^{-6}$
$0.6$	1.43333	1.43446	$1.13524\times {10}^{-3}$	$1.43334$	$1.08922\times {10}^{-5}$
$0.7$	1.63232	1.63419	$1.87147\times {10}^{-3}$	$1.63233$	$1.81055\times {10}^{-5}$
$0.8$	1.89648	1.89952	$3.04087\times {10}^{-3}$	$1.89651$	$2.9567\times {10}^{-5}$
$0.9$	2.24791	2.25281	$4.90399\times {10}^{-3}$	$2.24796$	$4.78283\times {10}^{-5}$
$1.0$	2.71828	2.72617	$7.88713\times {10}^{-3}$	$2.71836$	$7.70601\times {10}^{-5}$

| Show Table

DownLoad: CSV

Table 4. Numerical result for approximate solution of

${\omega }_{2}\left(\tau \right)$ in Problem 2.

$\tau$	${\omega }_{2}\left(\tau \right)$	${\mu }_{2}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.1$ )	$\left\|{e}_{2}\left(\tau \right)\right\|$	${\mu }_{2}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.01$ )	$\left\|{e}_{2}\left(\tau \right)\right\|$
0.0	0	$5.99863\times {10}^{-31}$	$5.99863\times {10}^{-31}$	0	0
0.1	$-0.0009$	$-0.000853346$	$4.66538\times {10}^{-5}$	$-0.000899887$	$1.13186\times {10}^{-7}$
0.2	$-0.0064$	$-0.0063379$	$6.21025\times {10}^{-5}$	$-0.00639955$	$4.45865\times {10}^{-7}$
$0.3$	-0.0189	$-0.0187802$	$1.19801\times {10}^{-4}$	$-0.01889900$	$9.98298\times {10}^{-7}$
$0.4$	-0.0384	$-0.0382022$	$1.97763\times {10}^{-4}$	$-0.0383982$	$1.76662\times {10}^{-6}$
$0.5$	-0.0625	$-0.0622052$	$2.94843\times {10}^{-4}$	$-0.0624973$	$2.74383\times {10}^{-6}$
$0.6$	-0.0864	$-0.0859881$	$4.11936\times {10}^{-4}$	$-0.0863961$	$3.91802\times {10}^{-6}$
$0.7$	-0.1029	$-0.102353$	$5.46598\times {10}^{-4}$	$-0.102895$	$5.2695\times {10}^{-6}$
$0.8$	-0.1024	$-0.101704$	$6.95542\times {10}^{-4}$	$-0.102393$	$6.76623\times {10}^{-6}$
$0.9$	-0.0729	$-0.0720466$	$8.5344\times {10}^{-4}$	$-0.0728916$	$8.35631\times {10}^{-6}$
$1.0$	0	$0.00101164$	$1.01164\times {10}^{-3}$	$9.95549\times {10}^{-6}$	$9.95549\times {10}^{-6}$

| Show Table

DownLoad: CSV

Table 5. Comparison of maximum absolute error of Problem 2.

	${\mu }_{1}\left(\tau \right)$	${\mu }_{2}\left(\tau \right)$
CBSM $({\mathrm{\Lambda }}=0.1)$	$7.88\times {10}^{-3}$	$1.01\times {10}^{-3}$
CBSM $({\mathrm{\Lambda }}=0.01)$	$7.70\times {10}^{-5}$	$9.95\times {10}^{-6}$
^[27] (N=5)	$3.14\times {10}^{-2}$	$4.19\times {10}^{-5}$
^[27] (N=6)	$6.11\times {10}^{-4}$	$7.22\times {10}^{-6}$

| Show Table

DownLoad: CSV

Problem 3. Consider the following system ^[24,41]

${\omega }_{1}^{{\text{'}}{\text{'}}}\left(\tau \right)+\frac{5}{\tau }{\omega }_{1}^{{\text{'}}}\left(\tau \right)+8({e}^{{\omega }_{1}\left(\tau \right)}+2{e}^{-\frac{{\omega }_{2}\left(\tau \right)}{2}}) = 0,\\ {\omega }_{2}^{{\text{'}}{\text{'}}}\left(\tau \right)+\frac{3}{\tau }{\omega }_{2}^{{\text{'}}}\left(\tau \right)-8({e}^{\frac{{\omega }_{1}\left(\tau \right)}{2}}+{e}^{-{\omega }_{2}\left(\tau \right)}) = 0,$

(30)

subject to

${\omega }_{1}\left(0\right) = 1-2{\mathrm{ln}}\left(2\right),{\omega }_{1}^{{\text{'}}}\left(0\right) = 0,\\{\omega }_{2}\left(0\right) = 1+2{\mathrm{ln}}\left(2\right),{\omega }_{2}^{{\text{'}}}\left(0\right) = 0,$

(31)

where the exact solutions are

${\omega }_{1}\left(\tau \right) = 1-2{\mathrm{ln}}({\tau }^{2}+2)$

and

${\omega }_{2}\left(\tau \right) = 1+2{\mathrm{ln}}({\tau }^{2}+2).$

We depicted our numerical and exact solutions with different values of ${\mathrm{\Lambda }}$ in Figure 2. The absolute errors between exact and numerical results are reported in Tables 6 and 7. Table 8 compares the maximum absolute error of CBSM with those of the method in ^[41]. It appears that our findings are superior to the outcomes presented in ^[41]. The CPU time for this problem, with ${\mathrm{\Lambda }} = 0.1$ and $0.01$ , are 0.0156s and 0.0781s respectively.

Figure 2. Absolute error functions for Problem 3.

DownLoad: Full-Size Img PowerPoint

Table 6. Numerical result for approximate solution of

${\omega }_{1}\left(\tau \right)$ in Problem 3.

$\tau$	${\omega }_{1}\left(\tau \right)$	${\mu }_{1}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.1$ )	$\left\|{e}_{1}\left(\tau \right)\right\|$	${\mu }_{1}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.01$ )	$\left\|{e}_{1}\left(\tau \right)\right\|$
$0.0$	$-0.386294$	$-0.386294$	$2.22045\times {10}^{-16}$	$-0.386294$	$-2.22045\times {10}^{-16}$
$0.1$	$-0.396269$	$-0.396257$	$1.28893\times {10}^{-5}$	$-0.396269$	$4.13534\times {10}^{-8}$
$0.2$	$-0.4259$	$-0.425878$	$2.11827\times {10}^{-5}$	$-0.425899$	$1.5319\times {10}^{-7}$
$0.3$	$-0.474328$	$-0.474293$	$3.503\times {10}^{-5}$	$-0.474328$	$3.07878\times {10}^{-7}$
$0.4$	$-0.540216$	$-0.540166$	$5.02533\times {10}^{-5}$	$-0.540216$	$4.67063\times {10}^{-7}$
$0.5$	$-0.62186$	$-0.621799$	$6.18095\times {10}^{-5}$	$-0.62186$	$5.91857\times {10}^{-7}$
$0.6$	$-0.717323$	$-0.717256$	$6.68475\times {10}^{-5}$	$-0.717323$	$6.51126\times {10}^{-7}$
$0.7$	$-0.824565$	$-0.824502$	$6.3593\times {10}^{-5}$	$-0.824565$	$6.26415\times {10}^{-7}$
$0.8$	$-0.941558$	$-0.941506$	$5.16047\times {10}^{-5}$	$-0.941557$	$5.13027\times {10}^{-7}$
$0.9$	$-1.06637$	$-1.06634$	$3.15988\times {10}^{-5}$	$-1.06637$	$3.18094\times {10}^{-7}$
$1.0$	$-1.19722$	$-1.19722$	$5.10219\times {10}^{-6}$	$-1.19722$	$5.69956\times {10}^{-8}$

| Show Table

DownLoad: CSV

Table 7. Numerical result for approximate solution of

${\omega }_{2}\left(\tau \right)$ in Problem 3.

$\tau$	${\omega }_{2}\left(\tau \right)$	${\mu }_{2}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.1$ )	$\left\|{e}_{2}\left(\tau \right)\right\|$	${\mu }_{2}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.01$ )	$\left\|{e}_{2}\left(\tau \right)\right\|$
$0.0$	2.38629	2.38629	$2.22045\times {10}^{-16}$	2.38629	$2.22045\times {10}^{-16}$
$0.1$	2.39627	2.39625	$1.47619\times {10}^{-5}$	2.39627	$6.20966\times {10}^{-8}$
$0.2$	2.4259	2.42587	$3.17866\times {10}^{-5}$	2.42599	$2.32511\times {10}^{-7}$
$0.3$	2.47433	2.47427	$5.42035\times {10}^{-5}$	2.47433	$4.75075\times {10}^{-7}$
$0.4$	2.54022	2.54014	$7.94034\times {10}^{-5}$	2.54022	$7.3881\times {10}^{-7}$
$0.5$	2.62186	2.62176	$1.01507\times {10}^{-4}$	2.62186	$9.70554\times {10}^{-7}$
$0.6$	2.71732	2.71721	$1.16022\times {10}^{-4}$	2.71732	$1.12546\times {10}^{-6}$
$0.7$	2.82457	2.82445	$1.1998\times {10}^{-4}$	2.82456	$1.17361\times {10}^{-6}$
$0.8$	2.94156	2.94145	$1.121\times {10}^{-4}$	2.94156	$1.10206\times {10}^{-6}$
$0.9$	3.06637	3.06628	$9.26097\times {10}^{-5}$	3.06637	$9.13128\times {10}^{-7}$
$1.0$	3.19722	3.19716	$6.2868\times {10}^{-5}$	3.19722	$6.20546\times {10}^{-7}$

| Show Table

DownLoad: CSV

Table 8. Comparison of maximum absolute error of Problem 3.

	${\mu }_{1}\left(\tau \right)$	${\mu }_{2}\left(\tau \right)$
CBSM $({\mathrm{\Lambda }}=0.1)$	$6.68\times {10}^{-5}$	$1.19\times {10}^{-4}$
CBSM $({\mathrm{\Lambda }}=0.01)$	$6.51\times {10}^{-7}$	$1.17\times {10}^{-6}$
^[41] (j=3)	$1.47\times {10}^{-3}$	$1.64\times {10}^{-3}$
^[41] (j=4)	$3.67\times {10}^{-4}$	$4.10\times {10}^{-4}$

| Show Table

DownLoad: CSV

Problem 4. Consider the following system of LEE ^{[24,27,40,41]}

${\omega }_{1}^{{\text{'}}{\text{'}}}\left(\tau \right)+\frac{1}{\tau }{\omega }_{1}^{{\text{'}}}\left(\tau \right)-{\omega }_{2}^{3}\left(\tau \right)({\omega }_{1}^{2}+1) = 0,\\{\omega }_{2}^{{\text{'}}{\text{'}}}\left(\tau \right)+\frac{3}{\tau }{\omega }_{2}^{{\text{'}}}\left(\tau \right)+{\omega }_{2}^{5}\left(\tau \right)({\omega }_{1}^{2}+3) = 0,$

(32)

subject to

${\omega }_{1}\left(0\right) = 1,{{\omega }^{{\text{'}}}}_{1}\left(0\right) = 0,\\{\omega }_{2}\left(0\right) = 1,{{\omega }^{{\text{'}}}}_{2}\left(0\right) = 0,$

(33)

where the exact solutions are

${\omega }_{1}\left(\tau \right) = \sqrt{{1+\tau }^{2}}$

and

${\omega }_{2}\left(\tau \right) = \frac{1}{\sqrt{{1+\tau }^{2}}}.$

The achieved numerical results with ${\mathrm{\Lambda }} = 0.1$ and $0.01$ are exposed in Tables 9 and 10. Table 11 shows a comparison between the maximum absolute error of CBSM and that of the approaches discussed in ^[27,40,41]. Based on the results, it seems that our research outperforms the results reported in ^[27,40,41]. Absolute errors for different values of ${\mathrm{\Lambda }}$ are exposed in Figure 3. The CPU time for this problem, with ${\mathrm{\Lambda }} = 0.1$ and $0.01$ , are 0.0156s and 0.0781s respectively.

Table 9. Numerical result for approximate solution of

${\omega }_{1}\left(\tau \right)$ in Problem 4.

$\tau$	${\omega }_{1}\left(\tau \right)$	${\mu }_{1}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.1$ )	$\left\|{e}_{1}\left(\tau \right)\right\|$	${\mu }_{1}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.01$ )	$\left\|{e}_{1}\left(\tau \right)\right\|$
$0.0$	1	1	0	1	0
$0.1$	1.004988	1.004978	$9.41481\times {10}^{-6}$	1.004987	$6.21775\times {10}^{-8}$
$0.2$	1.019804	1.019776	$2.80913\times {10}^{-5}$	1.019804	$2.29506\times {10}^{-7}$
$0.3$	1.044031	1.043979	$5.15984\times {10}^{-5}$	1.04403	$4.61458\times {10}^{-7}$
$0.4$	1.077033	1.076957	$7.60906\times {10}^{-5}$	1.077032	$7.07716\times {10}^{-7}$
$0.5$	1.118034	1.117936	$9.76652\times {10}^{-5}$	1.118033	$9.25606\times {10}^{-7}$
$0.6$	1.16619	1.166076	$1.13971\times {10}^{-4}$	1.166189	$1.09013\times {10}^{-6}$
$0.7$	1.220656	1.220531	$1.24491\times {10}^{-4}$	1.220654	$1.19564\times {10}^{-6}$
$0.8$	1.280625	1.280495	$1.30176\times {10}^{-4}$	1.280624	$1.25163\times {10}^{-6}$
$0.9$	1.345362	1.34523	$1.32854\times {10}^{-4}$	1.345361	$1.27653\times {10}^{-6}$
$1.0$	1.414214	1.414079	$1.34662\times {10}^{-4}$	1.414212	$1.29187\times {10}^{-6}$

| Show Table

DownLoad: CSV

Table 10. Numerical result for approximate solution of

${\omega }_{2}\left(\tau \right)$ in Problem 4.

$\tau$	${\omega }_{2}\left(\tau \right)$	${\mu }_{2}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.1$ )	$\left\|{e}_{2}\left(\tau \right)\right\|$	${\mu }_{2}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.01$ )	$\left\|{e}_{2}\left(\tau \right)\right\|$
$0.0$	1	1	0	1	$2.22045\times {10}^{-16}$
$0.1$	0.995037	0.995059	$2.1554\times {10}^{-5}$	0.995037	$9.03189\times {10}^{-8}$
$0.2$	0.980581	0.980623	$4.20048\times {10}^{-5}$	0.980581	$3.08008\times {10}^{-7}$
$0.3$	0.957826	0.957887	$6.03609\times {10}^{-5}$	0.957827	$5.34936\times {10}^{-7}$
$0.4$	0.928477	0.928545	$6.87324\times {10}^{-5}$	0.928477	$6.48155\times {10}^{-7}$
$0.5$	0.894427	0.894487	$5.93414\times {10}^{-5}$	0.894428	$5.76216\times {10}^{-7}$
$0.6$	0.857493	0.857525	$3.17569\times {10}^{-5}$	0.857493	$3.15637\times {10}^{-7}$
$0.7$	0.819232	0.819223	$9.41889\times {10}^{-6}$	0.819232	$8.61722\times {10}^{-8}$
$0.8$	0.780869	0.780811	$5.74405\times {10}^{-5}$	0.780868	$5.60283\times {10}^{-7}$
$0.9$	0.743294	0.743188	$1.05817\times {10}^{-4}$	0.743293	$1.04065\times {10}^{-6}$
$1.0$	0.707107	0.706957	$1.4962\times {10}^{-4}$	0.707105	$1.4772\times {10}^{-6}$

| Show Table

DownLoad: CSV

Table 11. Comparison of maximum absolute error of Problem 4.

	${\mu }_{1}\left(\tau \right)$	${\mu }_{2}\left(\tau \right)$
CBSM $({\mathrm{\Lambda }}=0.1)$	$1.34\times {10}^{-4}$	$1.49\times {10}^{-4}$
CBSM $({\mathrm{\Lambda }}=0.01)$	$1.29\times {10}^{-6}$	$1.47\times {10}^{-6}$
^[27] (N=4)	$6.44\times {10}^{-4}$	$8.87\times {10}^{-4}$
^[27] (N=5)	$7.46\times {10}^{-5}$	$6.08\times {10}^{-5}$
^[40] (n=4)	$4.86\times {10}^{-4}$	$1.45\times {10}^{-3}$
^[41] (j=3)	$2.76\times {10}^{-5}$	$1.31\times {10}^{-4}$
^[41] (j=4)	$6.87\times {10}^{-6}$	$3.28\times {10}^{-5}$

| Show Table

DownLoad: CSV

Figure 3. Absolute error functions for Problem 4.

DownLoad: Full-Size Img PowerPoint

Problem 5. Consider the following system of LEE ^[24,40]

${\omega }_{1}^{{\text{'}}{\text{'}}}+\frac{8}{\tau }{\omega }_{1}^{{\text{'}}}\left(\tau \right)+(18{\omega }_{1}\left(\tau \right)-4{\omega }_{1}\left(\tau \right){\mathrm{ln}}{\omega }_{2}\left(\tau \right)) = 0,\\{\omega }_{2}^{{\text{'}}{\text{'}}}\left(\tau \right)+\frac{4}{\tau }{\omega }_{2}^{{\text{'}}}\left(\tau \right)+(4{\omega }_{2}\left(\tau \right){\mathrm{ln}}{\omega }_{1}\left(\tau \right)-10{\omega }_{2}\left(\tau \right)) = 0,$

(34)

subject to

${\omega }_{1}\left(0\right) = 1,{\omega }_{1}^{{\text{'}}}\left(0\right) = 0, \\ {\omega }_{2}\left(0\right) = 1,{\omega }_{2}^{{\text{'}}}\left(0\right) = 0,$

(35)

where the exact solutions are ${\omega }_{1}\left(\tau \right) = {e}^{-{\tau }^{2}}$ and ${\omega }_{2}\left(\tau \right) = {e}^{{\tau }^{2}}$ .

Figure 4 represents the plot of our numerical and exact solutions for Problem 5 with different values of ${\mathrm{\Lambda }}$ . The absolute errors for ${\mathrm{\Lambda }} = 0.1$ and $0.01$ are presented in Tables 12 and 13. Table 14 displays a comparison between the maximum absolute error of CBSM and that of the approaches discussed in ^[40,41]. Our results indicate that they are better than the results reported in ^[40,41]. The CPU time for this problem, with ${\mathrm{\Lambda }} = 0.1$ and $0.01$ , are 0.0156s and 0.0625s respectively. From all these tables and figures, we can observe that our numerical results are in good agreement with the exact ones. Moreover, it appears from our findings that the cubic B-spline method exhibits more accurate results than some existing methods.

Figure 4. Absolute error functions for Problem 5.

DownLoad: Full-Size Img PowerPoint

Table 12. Numerical result for approximate solution of

${\omega }_{1}\left(\tau \right)$ in Problem 5.

$\tau$	${\omega }_{1}\left(\tau \right)$	${\mu }_{1}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.1$ )	$\left\|{e}_{1}\left(\tau \right)\right\|$	${\mu }_{1}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.01$ )	$\left\|{e}_{1}\left(\tau \right)\right\|$
$0.0$	1	1	0	1	0
$0.1$	0.99005	0.990073	$2.29673\times {10}^{-5}$	0.99005	$5.53154\times {10}^{-8}$
$0.2$	0.960789	0.960817	$2.7658"\times {10}^{-5}$	0.96079	$2.03451\times {10}^{-7}$
$0.3$	0.913931	0.913979	$4.76609\times {10}^{-5}$	0.913932	$4.05169\times {10}^{-7}$
$0.4$	0.852144	0.85221	$6.64108\times {10}^{-5}$	0.852144	$6.04412\times {10}^{-7}$
$0.5$	0.778801	0.778879	$7.81482\times {10}^{-5}$	0.778802	$7.42993\times {10}^{-7}$
$0.6$	0.697676	0.697756	$7.9594\times {10}^{-5}$	0.697677	$7.74528\times {10}^{-7}$
$0.7$	0.612626	0.612695	$6.83427\times {10}^{-5}$	0.612627	$6.75274\times {10}^{-7}$
$0.8$	0.527292	0.527337	$4.49264\times {10}^{-5}$	0.527293	$4.49447\times {10}^{-7}$
$0.9$	0.444858	0.444871	$1.24858\times {10}^{-5}$	0.444858	$1.28099\times {10}^{-7}$
$1.0$	0.367879	0.367856	$2.39084\times {10}^{-5}$	0.367879	$2.37741\times {10}^{-7}$

| Show Table

DownLoad: CSV

Table 13. Numerical result for approximate solution of

${\omega }_{2}\left(\tau \right)$ in Problem 5.

$\tau$	${\omega }_{2}\left(\tau \right)$	${\mu }_{2}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.1$ )	$\left\|{e}_{2}\left(\tau \right)\right\|$	${\mu }_{2}\left(\tau \right)$ ( ${\mathrm{\Lambda }}=0.01$ )	$\left\|{e}_{2}\left(\tau \right)\right\|$
$0.0$	1	1	0	1	0
$0.1$	1.01005	1.010078	$2.82181\times {10}^{-5}$	1.01005	$1.0345\times {10}^{-7}$
$0.2$	1.040811	1.040871	$6.04278\times {10}^{-5}$	1.040811	$4.35512\times {10}^{-7}$
$0.3$	1.094174	1.094299	$1.24976\times {10}^{-4}$	1.094175	$1.0806\times {10}^{-6}$
$0.4$	1.173511	1.173751	$2.40286\times {10}^{-4}$	1.173513	$2.19844\times {10}^{-6}$
$0.5$	1.284025	1.284458	$4.32197\times {10}^{-4}$	1.284029	$4.06313\times {10}^{-6}$
$0.6$	1.433329	1.434077	$7.47197\times {10}^{-4}$	1.433337	$7.12898\times {10}^{-6}$
$0.7$	1.632316	1.633578	$1.26164\times {10}^{-3}$	1.632328	$1.21425\times {10}^{-5}$
$0.8$	1.896481	1.898582	$2.1011\times {10}^{-3}$	1.896501	$1.21425\times {10}^{-5}$
$0.9$	2.247908	2.251381	$3.47326\times {10}^{-3}$	2.247942	$3.3724\times {10}^{-5}$
$1.0$	2.718282	2.724006	$5.72394\times {10}^{-3}$	2.718338	$5.56973\times {10}^{-5}$

| Show Table

DownLoad: CSV

Table 14. Comparison of maximum absolute error of Problem 5.

	${\mu }_{1}\left(\tau \right)$	${\mu }_{2}\left(\tau \right)$
CBSM $({\mathrm{\Lambda }}=0.1)$	$7.95\times {10}^{-5}$	$5.72\times {10}^{-3}$
CBSM $({\mathrm{\Lambda }}=0.01)$	$7.74\times {10}^{-7}$	$5.56\times {10}^{-5}$
^[40] (n=4)	$2.99\times {10}^{-3}$	$9.00\times {10}^{-3}$
^[41] (j=3)	$4.54\times {10}^{-4}$	$5.05\times {10}^{-4}$
^[41] (j=4)	$1.88\times {10}^{-4}$	$1.26\times {10}^{-4}$

| Show Table

DownLoad: CSV

6. Conclusions

The system of Lane-Emden type equations describes a variety of phenomena in theoretical physics, star structure, and astrophysics. In this study, we introduce and examine the use of the cubic B-spline method for studying the solution of singular and nonlinear systems of Lane-Emden equations. To address the singularity that occurs at τ = 0, we use L'Hôpital's rule. We also evaluate the accuracy and validity of the proposed technique, demonstrating its success in solving the considered system. The presented test problems have shown the simplicity and applicability of the proposed method. We provide tabular and graphical representations to confirm its effectiveness, observing that our numerical solutions are in good agreement with the exact solutions. It is observed that our numerical solutions are in good agreement with the exact ones. Furthermore, we show that by decreasing the mesh size, the numerical results converge to the analytical solution, which confirms the convergence of the algorithm. It is noteworthy that the CPU time of the proposed method for each evaluated problem is under 1 second.

Acknowledgments

The authors express their gratitude to the dear referees, who wish to remain anonymous and the editor for their helpful suggestions, which improved the final version of this paper.

Conflict of interest

The authors declare no conflicts of interest.

References

[1]	G. N. Kouziokas, A new W-SVM kernel combining PSO-neural network transformed vector and Bayesian optimized SVM in GDP forecasting, Eng. Appl. Artif. Intel., 92 (2020), 103650. https://doi.org/10.1016/j.engappai.2020.103650 doi: 10.1016/j.engappai.2020.103650
[2]	O. Cepni, I. E. Güney, N. R. Swanson, Nowcasting and forecasting GDP in emerging markets using global financial and macroeconomic diffusion indexes, Int. J. Forecasting, 35 (2019), 555–572. https://doi.org/10.1016/j.ijforecast.2018.10.008 doi: 10.1016/j.ijforecast.2018.10.008
[3]	T. E. Simos, V. N. Katsikis, S. D. Mourtas, P. S. Stanimirović, Unique non-negative definite solution of the time-varying algebraic Riccati equations with applications to stabilization of LTV systems, Math. Comput. Simulat., 202 (2022), 164–180. https://doi.org/10.1016/j.matcom.2022.05.033 doi: 10.1016/j.matcom.2022.05.033
[4]	S. D. Mourtas, V. N. Katsikis, C. Kasimis, Feedback control systems stabilization using a bio-inspired neural network, EAI Endorsed Trans. AI Robotics, 1 (2022), 1–13. https://doi.org/10.4108/airo.v1i.17 doi: 10.4108/airo.v1i.17
[5]	N. Premalatha, A. V. Arasu, Prediction of solar radiation for solar systems by using ANN models with different back propagation algorithms, J. Appl. Res. Technol., 14 (2016), 206–214. https://doi.org/10.1016/j.jart.2016.05.001 doi: 10.1016/j.jart.2016.05.001
[6]	S. X. Lv, L. Wang, Multivariate wind speed forecasting based on multi-objective feature selection approach and hybrid deep learning model, Energy, 263 (2023), 126100. https://doi.org/10.1016/j.energy.2022.126100 doi: 10.1016/j.energy.2022.126100
[7]	C. Huang, X. Jia, Z. Zhang, A modified back propagation artificial neural network model based on genetic algorithm to predict the flow behavior of 5754 aluminum alloy, Materials, 11 (2018), 855. https://doi.org/10.3390/ma11050855 doi: 10.3390/ma11050855
[8]	S. Gayathri, A. K. Krishna, V. P. Gopi, P. Palanisamy, Automated binary and multiclass classification of diabetic retinopathy using Haralick and multiresolution features, IEEE Access, 8 (2020), 57497–57504. https://doi.org/10.1109/ACCESS.2020.2979753
[9]	L. Chen, Z. Huang, Y. Li, N. Zeng, M. Liu, A. Peng, et al., Weight and structure determination neural network aided with double pseudoinversion for diagnosis of flat foot, IEEE Access, 7 (2019), 33001–33008. https://doi.org/10.1109/ACCESS.2019.2903634 doi: 10.1109/ACCESS.2019.2903634
[10]	M. R. Daliri, A hybrid automatic system for the diagnosis of lung cancer based on genetic algorithm and fuzzy extreme learning machines, J. Medical Syst., 36 (2012), 1001–1005. https://doi.org/10.1007/s10916-011-9806-y doi: 10.1007/s10916-011-9806-y
[11]	R. J. S. Raj, S. J. Shobana, I. V. Pustokhina, D. A. Pustokhin, D. Gupta, K. Shankar, Optimal feature selection-based medical image classification using deep learning model in internet of medical things, IEEE Access, 8 (2020), 58006–58017. https://doi.org/10.1109/ACCESS.2020.2981337 doi: 10.1109/ACCESS.2020.2981337
[12]	S. D. Mourtas, V. N. Katsikis, Exploiting the Black-Litterman framework through error-correction neural networks, Neurocomputing, 498 (2022), 43–58. https://doi.org/10.1016/j.neucom.2022.05.036 doi: 10.1016/j.neucom.2022.05.036
[13]	S. D. Mourtas, A weights direct determination neuronet for time-series with applications in the industrial indices of the federal reserve bank of St. Louis, J. Forecasting, 14 (2022), 1512–1524. https://doi.org/10.1002/for.2874 doi: 10.1002/for.2874
[14]	S. X. Lv, L. Peng, H. Hu, L. Wang, Effective machine learning model combination based on selective ensemble strategy for time series forecasting, Inf. Sci., 612 (2022), 994–1023. https://doi.org/10.1016/j.ins.2022.09.002 doi: 10.1016/j.ins.2022.09.002
[15]	Y. Zhang, D. Guo, Z. Luo, K. Zhai, H. Tan, CP-activated WASD neuronet approach to Asian population prediction with abundant experimental verification, Neurocomputing, 198 (2016), 48–57. https://doi.org/10.1016/j.neucom.2015.12.111 doi: 10.1016/j.neucom.2015.12.111
[16]	Y. Zhang, Z. Xue, M. Xiao, Y. Ling, C. Ye, Ten-Quarter Projection for Spanish Central Government Debt via WASD Neuronet, In: International Conference on Neural Information Processing, Springer, 2017. 893–902.
[17]	F. Groes, P. Kircher, I. Manovskii, The U-shapes of occupational mobility, Rev. Econ. Stud., 82 (2015), 659–692. https://doi.org/10.1093/restud/rdu037 doi: 10.1093/restud/rdu037
[18]	I. N. Generalao, Measuring the telework potential of jobs: Evidence from the international standard classification of occupations, Philippine Rev. Econ., 58 (2021), 92–127. https://doi.org/10.37907/5erp1202jd doi: 10.37907/5erp1202jd
[19]	D. Lagios, S. D. Mourtas, P. Zervas, G. Tzimas, A weights direct determination neural network for international standard classification of occupations, Mathematics, 11 (2023), 629. https://doi.org/10.3390/math11030629 doi: 10.3390/math11030629
[20]	J. Garnitz, R. Lehmann, K. Wohlrabe, Forecasting GDP all over the world using leading indicators based on comprehensive survey data, Appl. Econ., 51 (2019), 5802–5816. https://doi.org/10.1080/00036846.2019.1624915 doi: 10.1080/00036846.2019.1624915
[21]	M. Marcellino, M. Porqueddu, F. Venditti, Short-term GDP forecasting with a mixed-frequency dynamic factor model with stochastic volatility, J. Bus. Econ. Stat., 34 (2016), 118–127. https://doi.org/10.1080/07350015.2015.1006773 doi: 10.1080/07350015.2015.1006773
[22]	C. Liu, W. Xie, T. Lao, Y. Yao, J. Zhang, Application of a novel grey forecasting model with time power term to predict China's GDP, Grey Syst. Theory Appl., 11 (2021), 343–357. https://doi.org/10.1108/GS-05-2020-0065 doi: 10.1108/GS-05-2020-0065
[23]	J. Yoon, Forecasting of real GDP growth using machine learning models: Gradient boosting and random forest approach, Comput. Econ., 57 (2021), 247–265. https://doi.org/10.1007/s10614-020-10054-w doi: 10.1007/s10614-020-10054-w
[24]	H. H. Kim, N. R. Swanson, Methods for backcasting, nowcasting and forecasting using factor-MIDAS: With an application to Korean GDP, J. Forecasting, 37 (2018), 281–302. https://doi.org/10.1002/for.2499 doi: 10.1002/for.2499
[25]	A. Richardson, T. van Florenstein Mulder, T. Vehbi, Nowcasting GDP using machine-learning algorithms: A real-time assessment, Int. J. Forecasting, 37 (2021), 941–948. https://doi.org/10.1016/j.ijforecast.2020.10.005 doi: 10.1016/j.ijforecast.2020.10.005
[26]	Y. Zhang, D. Chen, C. Ye, Deep Neural Networks: WASD Neuronet Models, Algorithms, and Applications, CRC Press: Boca Raton, FL, USA, 2019.
[27]	T. E. Simos, V. N. Katsikis, S. D. Mourtas, A multi-input with multi-function activated weights and structure determination neuronet for classification problems and applications in firm fraud and loan approval, Appl. Soft Comput., 127 (2022), 109351. https://doi.org/10.1016/j.asoc.2022.109351 doi: 10.1016/j.asoc.2022.109351
[28]	T. E. Simos, S. D. Mourtas, V. N. Katsikis, Time-varying Black-Litterman portfolio optimization using a bio-inspired approach and neuronets, Appl. Soft Comput., 112 (2021), 107767. https://doi.org/10.1016/j.asoc.2021.107767 doi: 10.1016/j.asoc.2021.107767
[29]	Y. Zhang, X. Yu, L. Xiao, W. Li, Z. Fan, W. Zhang, Weights and structure determination of articial neuronets, In: Self-Organization: Theories and Methods, New York, NY, USA: Nova Science, 2013.
[30]	G. P. Zhang, D. M. Kline, Quarterly time-series forecasting with neural networks, IEEE T. Neur. Network., 18 (2007), 1800–1814. https://doi.org/10.1109/TNN.2007.896859 doi: 10.1109/TNN.2007.896859
[31]	A. Tharwat, Classification assessment methods, Appl. Comput. Inf., 17 (2020), 168–192. https://doi.org/10.1016/j.aci.2018.08.003 doi: 10.1016/j.aci.2018.08.003

This article has been cited by:

1.	Osama Ala'yed, Ahmad Qazza, Rania Saadeh, Osama Alkhazaleh, A quintic B-spline technique for a system of Lane-Emden equations arising in theoretical physical applications, 2024, 9, 2473-6988, 4665, 10.3934/math.2024225
2.	Oday Hazaimah, Polytropic Dynamical Systems with Time Singularity, 2024, 2581-8147, 721, 10.34198/ejms.14424.721746
3.	Ahmad El-Ajou, Mohammed Shqair, Ibrahim Ghabar, Aliaa Burqan, Rania Saadeh, A solution for the neutron diffusion equation in the spherical and hemispherical reactors using the residual power series, 2023, 11, 2296-424X, 10.3389/fphy.2023.1229142
4.	Richard Olu Awonusika, Oluwaseun Biodun Onuoha, Analytical method for systems of nonlinear singular boundary value problems, 2024, 11, 26668181, 100762, 10.1016/j.padiff.2024.100762
5.	Ajjanna Roja, Rania Saadeh, Raman Kumar, Ahmad Qazza, Umair Khan, Anuar Ishak, El-Sayed M. Sherif, Ioan Pop, Ramification of Hall effects in a non-Newtonian model past an inclined microchannel with slip and convective boundary conditions, 2024, 34, 1617-8106, 10.1515/arh-2024-0010
6.	Saeid Abbasbandy, A hybrid quantum-spectral-successive linearization method for general Lane–Emden type equations, 2024, 1598-5865, 10.1007/s12190-024-02290-2
7.	Rania Saadeh, Iqbal M. Batiha, Ahmad Qazza, Iqbal H. Jebri, Ali Elrashidi, Shaher Momani, 2024, Controlling DC Motor Speed with FoPID-Controllers, 979-8-3315-4001-2, 1, 10.1109/ACIT62805.2024.10877226
8.	Ahmad Qazza, Issam Bendib, Raed Hatamleh, Rania Saadeh, Adel Ouannas, Dynamics of the Gierer–Meinhardt reaction–diffusion system: Insights into finite-time stability and control strategies, 2025, 14, 26668181, 101142, 10.1016/j.padiff.2025.101142

Reader Comments

Your name:*

Email:*
© 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)