Issa Cherif Geraldo
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Issa Cherif Geraldo has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
- Tchilabalo Abozou Kpanzou
Laboratoire de Modélisation Mathématique et d’Analyse Statistique Décisionnelle (LaMMASD), Département de Mathématiques, Faculté des Sciences et Techniques, Université de Kara, Kara, Togo
6 shared publications - Edoh Katchekpele
Département de Mathématiques Laboratoire de Modélisation Mathématique et d'Analyse Statistique Décisionnelle (LaMMASD) Université de Kara BP 404 Kara TOGO
5 shared publications - Amiya Ranjan Mohanty
Indian Institute of Technology Kharagpur
2 shared publications - Jean-Philippe Cassar
Univ. Lille, CNRS, UMR 9189 - CRIStAL - Centre de Recherche en Informatique Signal et Automatique de Lille, F-59000 Lille, France
2 shared publications - Komi Midzodzi Pekpe
LAGIS UMR CNRS 8219, LILLE 1 University Villeneuve d'Ascq 59655, France
2 shared publications - Kevin Paumel
CEA DES IRESNE, Cadarache, Saint Paul lez Durance, France
1 shared publication
Latest Publications
- 1 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
On the consistency of some constrained maximum likelihood estimator used in crash data modelling
A propos de la consistance d’un estimateur du maximum de vraisemblance utilisé dans la modélisation de données d’accidents: L'ensemble des méthodes statistiques utilisées dans la modélisation de données nécessite la recherche de solutions optimales locales mais aussi l’estimation de la précision (écart-type) liée à ces solutions. Ces méthodes consistent à optimiser, par approximations itératives, la fonction de vraisemblance ou une version approchée. Classiquement, on utilise des versions adaptées de la méthode de Newton-Raphson ou des scores de Fisher. Du fait qu'elles nécessitent des inversions matricielles, ces méthodes peuvent être complexes à mettre en œuvre numériquement en grandes dimensions ou lorsque les matrices impliquées ne sont pas inversibles. Pour contourner ces difficultés, des procédures itératives ne nécessitant pas d’inversion matricielle telles que les algorithmes MM (Minorization-Maximization) ont été proposées et sont considérés comme pertinents pour les problèmes en grandes dimensions et pour certaines distributions discrètes multivariées. Parmi les nouvelles approches proposées dans le cadre de la modélisation en sécurité routière, figure un algorithme nommé algorithme cyclique itératif (CA). Cette thèse a un double objectif. Le premier est d'étudier l'algorithme CA des points de vue algorithmique et stochastique; le second est de généraliser l'algorithme cyclique itératif à des modèles plus complexes intégrant des distributions discrètes multivariées et de comparer la performance de l’algorithme CA généralisé à celle de ses compétiteurs.
- 1 April 2026
- Published by Agence Bibliographique de l'Enseignement Supérieur
On the consistency of some constrained maximum likelihood estimator used in crash data modelling
A propos de la consistance d’un estimateur du maximum de vraisemblance utilisé dans la modélisation de données d’accidents: L'ensemble des méthodes statistiques utilisées dans la modélisation de données nécessite la recherche de solutions optimales locales mais aussi l’estimation de la précision (écart-type) liée à ces solutions. Ces méthodes consistent à optimiser, par approximations itératives, la fonction de vraisemblance ou une version approchée. Classiquement, on utilise des versions adaptées de la méthode de Newton-Raphson ou des scores de Fisher. Du fait qu'elles nécessitent des inversions matricielles, ces méthodes peuvent être complexes à mettre en œuvre numériquement en grandes dimensions ou lorsque les matrices impliquées ne sont pas inversibles. Pour contourner ces difficultés, des procédures itératives ne nécessitant pas d’inversion matricielle telles que les algorithmes MM (Minorization-Maximization) ont été proposées et sont considérés comme pertinents pour les problèmes en grandes dimensions et pour certaines distributions discrètes multivariées. Parmi les nouvelles approches proposées dans le cadre de la modélisation en sécurité routière, figure un algorithme nommé algorithme cyclique itératif (CA). Cette thèse a un double objectif. Le premier est d'étudier l'algorithme CA des points de vue algorithmique et stochastique; le second est de généraliser l'algorithme cyclique itératif à des modèles plus complexes intégrant des distributions discrètes multivariées et de comparer la performance de l’algorithme CA généralisé à celle de ses compétiteurs.
- Open Access
- 9 January 2026
- Published by SCIK Publishing Corporation in European Journal of Statistics
- Vol. 6, 2
A Note on Local Maxima in Quadratic Transmuted Distributions Likelihoods
Transmutation is a widely used technique for generalizing probability distributions to improve data fitting. Its implementation often relies on maximum likelihood estimation, which reduces to a box-constrained numerical optimization problem. Despite this, many studies overlook the crucial role of the initial values required to start the optimization algorithm. In this paper, we demonstrate through two case studies on real data that improper parameter initialization can lead to convergence toward local maxima, ultimately resulting in biased estimates and incorrect conclusions. We show that the choice of starting values can significantly affect both the convergence behavior and the reliability of the final results. This study highlights the need for greater methodological rigor and increased awareness regarding parameter initialization in iterative estimation procedures, particularly within the context of transmuted distributions in order to avoid erroneous conclusions.
- 9 January 2026
- Published by SCIK Publishing Corporation in European Journal of Statistics
- Vol. 6, 2
Open AccessA Note on Local Maxima in Quadratic Transmuted Distributions Likelihoods
Transmutation is a widely used technique for generalizing probability distributions to improve data fitting. Its implementation often relies on maximum likelihood estimation, which reduces to a box-constrained numerical optimization problem. Despite this, many studies overlook the crucial role of the initial values required to start the optimization algorithm. In this paper, we demonstrate through two case studies on real data that improper parameter initialization can lead to convergence toward local maxima, ultimately resulting in biased estimates and incorrect conclusions. We show that the choice of starting values can significantly affect both the convergence behavior and the reliability of the final results. This study highlights the need for greater methodological rigor and increased awareness regarding parameter initialization in iterative estimation procedures, particularly within the context of transmuted distributions in order to avoid erroneous conclusions.
- Open Access
- 1 October 2025
- Published by MKD Publishing House in Malaya Journal of Matematik
- Vol. 13 (04) , 326-330
A note on a formula for quartic transmutation of probability distributions
In statistics, recent research has focused on developing new probability distributions that are better suited to modeling complex data than classical distributions. Such extensions typically involve applying mathematical transformations to a baseline cumulative distribution function (cdf), often introducing additional parameters. If these parameters are not properly defined, the resulting cdf may be ill-defined, potentially leading to erroneous conclusions in applications. In this note, we examine the formula for the quartic transmutation of probability distributions proposed by AL-Kadim [Journal of University of Babylon for Pure and Applied Sciences, 26(4):66–74, 2018]. We identify and correct an error in the original paper by specifying a new parameter space that guarantees the resulting cdf is well-defined.
- 1 October 2025
- Published by MKD Publishing House in Malaya Journal of Matematik
- Vol. 13 (04) , 326-330
Open AccessA note on a formula for quartic transmutation of probability distributions
In statistics, recent research has focused on developing new probability distributions that are better suited to modeling complex data than classical distributions. Such extensions typically involve applying mathematical transformations to a baseline cumulative distribution function (cdf), often introducing additional parameters. If these parameters are not properly defined, the resulting cdf may be ill-defined, potentially leading to erroneous conclusions in applications. In this note, we examine the formula for the quartic transmutation of probability distributions proposed by AL-Kadim [Journal of University of Babylon for Pure and Applied Sciences, 26(4):66–74, 2018]. We identify and correct an error in the original paper by specifying a new parameter space that guarantees the resulting cdf is well-defined.
- 11 July 2025
- Published by International Scientific Research Publications in Journal of Nonlinear Sciences and Applications
- Vol. 18 (04) , 250-258
Transmuted power function distribution revisited: simulation study and important lessons on starting values and local maxima
In this paper, we consider the transmuted power function distribution (TPFD), an important flexible distribution with applications in lifetime modelling. In past studies, maximum likelihood (ML) was recommended as the best method to use for estimating the parameters of the TPFD. We study the ML estimation problem from a novel computational angle by making a comprehensive study in R software using some of the best optimization algorithms. This study enables us to draw some important lessons on some of the computational aspects sometimes neglected in solving maximum likelihood estimation problems.
- 11 July 2025
- Published by International Scientific Research Publications in Journal of Nonlinear Sciences and Applications
- Vol. 18 (04) , 250-258
Transmuted power function distribution revisited: simulation study and important lessons on starting values and local maxima
In this paper, we consider the transmuted power function distribution (TPFD), an important flexible distribution with applications in lifetime modelling. In past studies, maximum likelihood (ML) was recommended as the best method to use for estimating the parameters of the TPFD. We study the ML estimation problem from a novel computational angle by making a comprehensive study in R software using some of the best optimization algorithms. This study enables us to draw some important lessons on some of the computational aspects sometimes neglected in solving maximum likelihood estimation problems.
- Open Access
Comparative analysis and practical applications of cubic transmutations for the Pareto distribution
Transmutation is a technique for extending classical probability distributions in order to give them more flexibility. In this paper, we are interested in cubic transmutations of the Pareto distribution. We establish a general formula that unifies existing cubic transmutations of the Pareto distribution and facilitates the derivation of new cubic transmutations that have not yet been explored in the literature. We also derive general formulas for the related mathematical properties. Finally, we perform a comparative analysis of the six transmutations existing in the literature using real-world data. The results obtained confirm the flexibility and effectiveness of cubic transmutations in modeling various types of data.
Open AccessComparative analysis and practical applications of cubic transmutations for the Pareto distribution
Transmutation is a technique for extending classical probability distributions in order to give them more flexibility. In this paper, we are interested in cubic transmutations of the Pareto distribution. We establish a general formula that unifies existing cubic transmutations of the Pareto distribution and facilitates the derivation of new cubic transmutations that have not yet been explored in the literature. We also derive general formulas for the related mathematical properties. Finally, we perform a comparative analysis of the six transmutations existing in the literature using real-world data. The results obtained confirm the flexibility and effectiveness of cubic transmutations in modeling various types of data.
- Open Access
- 1 January 2025
- Published by Wiley in International Journal of Mathematics and Mathematical Sciences
- Vol. 2025 (1)
Comparative Analysis and Practical Applications of Cubic Transmutations for the Pareto Distribution
Transmutation is a technique for extending classical probability distributions in order to give them more flexibility. In this paper, we are interested in cubic transmutations of the Pareto distribution. We establish a general formula that unifies the three existing cubic transmuted Pareto (CTP) distributions and facilitates the derivation of three new CTP distributions that have not yet been proposed in the literature. We also derive general formulas for the related mathematical properties and study parameter estimation via the maximum likelihood method. Finally, we perform a comparative analysis of the six CTP distributions using real‐world data. The results obtained confirm the flexibility and effectiveness of CTP distributions in modelling various types of data.
- 1 January 2025
- Published by Wiley in International Journal of Mathematics and Mathematical Sciences
- Vol. 2025 (1)
Open AccessComparative Analysis and Practical Applications of Cubic Transmutations for the Pareto Distribution
Transmutation is a technique for extending classical probability distributions in order to give them more flexibility. In this paper, we are interested in cubic transmutations of the Pareto distribution. We establish a general formula that unifies the three existing cubic transmuted Pareto (CTP) distributions and facilitates the derivation of three new CTP distributions that have not yet been proposed in the literature. We also derive general formulas for the related mathematical properties and study parameter estimation via the maximum likelihood method. Finally, we perform a comparative analysis of the six CTP distributions using real‐world data. The results obtained confirm the flexibility and effectiveness of CTP distributions in modelling various types of data.
- Open Access
- 26 September 2024
- Published by Vertex Academic Press in Gulf Journal of Mathematics
- Vol. 17 (2) , 226-242
Unveiling change-point detection with ARMA-GARCH models
This paper delves into the asymptotic theory of a Cumulative Sum (CUSUM)-type test designed for detecting change points in the variance of AutoRegressive Moving-Average models, incorporating Generalized Autoregressive Conditional Heteroscedasticity innovations. We demonstrate that, under the null hypothesis (no change), the CUSUM test statistic converges to the supremum of a standard Brownian bridge. Using a Monte Carlo simulation, we highlight the strong performance of our proposed test when compared to the methods by Song and Kang. Furthermore, we provide a real-world data analysis to showcase the practical application of our test.
- 26 September 2024
- Published by Vertex Academic Press in Gulf Journal of Mathematics
- Vol. 17 (2) , 226-242
Open AccessUnveiling change-point detection with ARMA-GARCH models
This paper delves into the asymptotic theory of a Cumulative Sum (CUSUM)-type test designed for detecting change points in the variance of AutoRegressive Moving-Average models, incorporating Generalized Autoregressive Conditional Heteroscedasticity innovations. We demonstrate that, under the null hypothesis (no change), the CUSUM test statistic converges to the supremum of a standard Brownian bridge. Using a Monte Carlo simulation, we highlight the strong performance of our proposed test when compared to the methods by Song and Kang. Furthermore, we provide a real-world data analysis to showcase the practical application of our test.
- Open Access
- 3 December 2023
- Published by Vertex Academic Press in Gulf Journal of Mathematics
- Vol. 15 (2) , 117-132
Exact computation of the asymptotic variance matrix for a tuple of multinomial distributions with dependent parameters
In this paper, we consider a statistical model consisting of an s-tuple of multinomial distributions whose parameters are dependent. This model is used in road safety to study the effect of a measure implemented on a set of s sites comprising a total of r accident severity levels and is depending on a parameter vector of 1+sr components under s equality constraints. The computation of the so important asymptotic variance matrix W requires some numerical matrix inversions which become very complicated for large values of s and r. To circumvent this difficulty, we use Schur complement for the exact analytic inversion and we propose a strategy for calculating W. We illustrate the effectiveness of our approach using simulated data.
- 3 December 2023
- Published by Vertex Academic Press in Gulf Journal of Mathematics
- Vol. 15 (2) , 117-132
Open AccessExact computation of the asymptotic variance matrix for a tuple of multinomial distributions with dependent parameters
In this paper, we consider a statistical model consisting of an s-tuple of multinomial distributions whose parameters are dependent. This model is used in road safety to study the effect of a measure implemented on a set of s sites comprising a total of r accident severity levels and is depending on a parameter vector of 1+sr components under s equality constraints. The computation of the so important asymptotic variance matrix W requires some numerical matrix inversions which become very complicated for large values of s and r. To circumvent this difficulty, we use Schur complement for the exact analytic inversion and we propose a strategy for calculating W. We illustrate the effectiveness of our approach using simulated data.
- Open Access
- 19 April 2023
- Published by Wiley in Journal of Applied Mathematics
- Vol. 2023, 1-10
A Fast and Efficient Estimation of the Parameters of a Model of Accident Frequencies via an MM Algorithm
In this paper, we consider a multivariate statistical model of accident frequencies having a variable number of parameters and whose parameters are dependent and subject to box constraints and linear equality constraints. We design a minorization-maximization (MM) algorithm and an accelerated MM algorithm to compute the maximum likelihood estimates of the parameters. We illustrate, through simulations, the performance of our proposed MM algorithm and its accelerated version by comparing them to Newton-Raphson (NR) and quasi-Newton algorithms. The results suggest that the MM algorithm and its accelerated version are better in terms of convergence proportion and, as the number of parameters increases, they are also better in terms of computation time.
- 19 April 2023
- Published by Wiley in Journal of Applied Mathematics
- Vol. 2023, 1-10
Open AccessA Fast and Efficient Estimation of the Parameters of a Model of Accident Frequencies via an MM Algorithm
In this paper, we consider a multivariate statistical model of accident frequencies having a variable number of parameters and whose parameters are dependent and subject to box constraints and linear equality constraints. We design a minorization-maximization (MM) algorithm and an accelerated MM algorithm to compute the maximum likelihood estimates of the parameters. We illustrate, through simulations, the performance of our proposed MM algorithm and its accelerated version by comparing them to Newton-Raphson (NR) and quasi-Newton algorithms. The results suggest that the MM algorithm and its accelerated version are better in terms of convergence proportion and, as the number of parameters increases, they are also better in terms of computation time.
- Open Access
- 26 October 2022
- Published by Hindawi Limited in Journal of Applied Mathematics
- Vol. 2022, 1-11
An Automated Profile-Likelihood-Based Algorithm for Fast Computation of the Maximum Likelihood Estimate in a Statistical Model for Crash Data
Numerical computation of maximum likelihood estimates (MLE) is one of the most common problems encountered in applied statistics. Even if there exist many algorithms considered as performing, they can suffer in some cases for one or many of the following criteria: global convergence (capacity of an algorithm to converge to the true unknown solution from all starting guesses), numerical stability (ascent property), implementation feasibility (for example, algorithms requiring matrix inversion cannot be implemented when the involved matrices are not invertible), low computation time, low computational complexity, and capacity to handle high dimensional problems. The reality is that, in practice, no algorithm is perfect, and for each problem, it is necessary to find the most performing of all existing algorithms or even develop new ones. In this paper, we consider the computing of the maximum likelihood estimate of the vector parameter of a statistical model of crash frequencies. We split the parameter vector, and we develop a new estimation algorithm using the profile likelihood principle. We provide an automatic starting guess for which convergence and numerical stability are guaranteed. We study the performance of our new algorithm on simulated data by comparing it to some of the most famous and modern optimization algorithms. The results suggest that our proposed algorithm outperforms these algorithms.
- 26 October 2022
- Published by Hindawi Limited in Journal of Applied Mathematics
- Vol. 2022, 1-11
Open AccessAn Automated Profile-Likelihood-Based Algorithm for Fast Computation of the Maximum Likelihood Estimate in a Statistical Model for Crash Data
Numerical computation of maximum likelihood estimates (MLE) is one of the most common problems encountered in applied statistics. Even if there exist many algorithms considered as performing, they can suffer in some cases for one or many of the following criteria: global convergence (capacity of an algorithm to converge to the true unknown solution from all starting guesses), numerical stability (ascent property), implementation feasibility (for example, algorithms requiring matrix inversion cannot be implemented when the involved matrices are not invertible), low computation time, low computational complexity, and capacity to handle high dimensional problems. The reality is that, in practice, no algorithm is perfect, and for each problem, it is necessary to find the most performing of all existing algorithms or even develop new ones. In this paper, we consider the computing of the maximum likelihood estimate of the vector parameter of a statistical model of crash frequencies. We split the parameter vector, and we develop a new estimation algorithm using the profile likelihood principle. We provide an automatic starting guess for which convergence and numerical stability are guaranteed. We study the performance of our new algorithm on simulated data by comparing it to some of the most famous and modern optimization algorithms. The results suggest that our proposed algorithm outperforms these algorithms.