Statistics, Optimization & Information Computing

Interior-point methods for monotone linear complementarity problems based on the new kernel function with applications to Control Tabular Adjustment problem

Lesaja Goran — 2025-01-19

We present a feasible kernel-based interior point method (IPM) to solve a monotone linear complementarity problem (LCP) which is based on an eligible kernel function with new logarithmic barrier term. The new kernel function defines new search direction and the neighborhood of the central path. We show the global convergence of the algorithm and derive the iteration bounds for short- and long-step versions of the algorithm.

We applied the method to solve a continuous Control Tabular Adjustment (CTA) problem which is an important Statistical Disclosure Limitation (SDL) model for protection of tabular data. Numerical results on a test example show that this algorithm is a viable option to the existing methods for solving continuous CTA problems. We also apply the algorithm to the set of randomly generated monotone LCPs showing that the initial implementation performs well on these instances of LCPs. However, this very limited numerical testing is done for illustration purposes only; an extensive numerical study is necessary to draw more definite conclusions on the behavior of the algorithm.

A Hybrid Direction Method for Linear Fractional Programming

Mohammed Amin Hakmi — 2025-01-02

In this article, we propose a new method for solving Linear Fractional Programming (LFP) problems with bounded variables. The proposed algorithm passes from a support feasible solution to a better one following the feasible direction proposed in [K. Djeloud, M. Bentobache and M. O. Bibi, A new method with hybrid direction for linear programming, Concurrency and Computation, Practice and Experience 33 (1), 2021]. Optimality and suboptimality criteria which allow to stop the algorithm when an optimal or suboptimal solution is achieved were stated and proved. Then, a new method called a Hybrid Direction Method (HDM) is described and a numerical example is given for illustration purpose. In order to compare our method to the classical approaches, we develop an implementation with the Matlab programming language. The obtained numerical results on solving 120 randomly generated LFP test problems show that HDM with long step rule is competitive with the primal simplex method and the interior-points method implemented in Matlab.

Optimal control analysis of a tuberculosis model with drug-resistant population

Cicik Alfiniyah — 2025-01-31

Tuberculosis (TB), caused by Mycobacterium tuberculosis, stands as one of the most infectious diseases globally, predominantly affecting the lungs (known as pulmonary tuberculosis). It manifests in two primary forms based on bacterial drug sensitivity: drug-sensitive TB (DS-TB) and drug-resistant TB (DR-TB). DS-TB remains susceptible to medication, whereas DR-TB has developed resistance. This study explores a mathematical model explaining the spread of tuberculosis within a drug-resistant population, proposing optimal control strategies to curb its dissemination through educational initiatives and enhancements in healthcare facilities. The stability analysis reveals that disease-free equilibrium points are locally asymptotically stable when R_0 < 1, while endemic equilibrium points prevail and are locally asymptotically stable if R_0 > 1. Additionally, sensitivity analysis identifies important parameters within the model. By using the Pontryagin Maximum Principle, control variables are integrated and numerically solved. Through simulations and cost assessments, we illustrate the efficacy of employing both control strategies concurrently, effectively reducing the populations susceptible to exposure, DS-TB, and DR-TB infections.

Uncertain Portfolio Optimization Problems: Systematic Review

Khalid Belabbes — 2024-11-26

This paper presents a literature review and analysis of Uncertain Portfolio Optimization Problems (UPOP), where security returns are described by uncertain variables due to a lack of historical data. To identify the major gaps in literature and offer perspectives for future research to address these limitations, this paper reviews more than 80 works that have shaped the field among foundational works and recent advancements in UPOP until 2024. We have presented the definitions and some comparisons between various mathematical risk measures to allow the decision-maker to choose which one is appropriate for their situation. In addition, some real features that marked literature are introduced. This has provided a number of enhancements that have been suggested as, artificial intelligence utilization, considering environmental constraints and, using other techniques to model asset returns as the uncertain random variables and employing dynamic and multi-period optimization methods.

Inverse Multi-Objective Optimization for Portfolio Allocation in Commercial Banks

Nagwa Albehery — 2025-01-09

Optimal portfolio allocation in commercial banks is a critical decision for financial institutions. This paper proposes a multi-objective linear programming model to address this challenge. To ensure the model's feasibility and efficiency, we employ a generalized inverse optimization approach, replacing regular optimality with Pareto optimality. We apply our proposed models to real data from Bank Misr, an Egyptian bank, during the finance year 2020/2021. The multi-objective model was solved using LINGO 19, while the inverse multi-objective model was solved using R programming. Our analysis of the results provides valuable insights into optimal portfolio distribution for commercial banks.

Bayesian Methods for Multi-Objective Optimization of Hybrid Numerical Filters in ECG Signal Processing for Accurate Arrhythmia Classification

Zakaria Khatar — 2024-12-05

This study introduces an innovative method for ECG signal processing that combines advanced filtering techniques, multi-objective Bayesian optimization, and a sophisticated deep learning architecture for classification. The methodology starts with Enhanced Empirical Mode Decomposition (EEMD) to break down the ECG signal into Intrinsic Mode Functions (IMFs). These IMFs undergo filtration through a series of Chebyshev Type II, Butterworth, Daubechies Wavelet, and Savitzky-Golay filters. To achieve optimal performance, a Bayesian multi-objective optimization strategy, augmented by reinforcement learning for dynamic weight adjustment and Gaussian process minimization, is utilized to fine-tune filter parameters. This process ensures maximum noise reduction while maintaining signal integrity. The optimized signals are then processed by an advanced deep learning architecture that includes parallel and residual connections, bidirectional GRU layers, and dense classification layers, enabling precise classification of cardiac conditions. The model's performance was rigorously tested across 12 different ECG leads, showing remarkable improvements in classification accuracy (ACC), sensitivity (SNS), and F1 score. Post-optimization results achieved impressive values of 99.24\% for ACC, 99.04\% for SNS, and 99.05\% for F1 score, demonstrating the significant enhancement in ECG signal analysis and diagnostic reliability provided by the proposed approach.

Multi-modal Stacked Ensemble Model for Breast Cancer Prognosis Prediction

Aminu Maigari — 2024-10-15

Breast cancer (BC) is a global health challenge that affects millions of women worldwide and leads to significant mortality. Recent advancements in next-generation sequencing technology have enabled comprehensive diagnosis and prognosis determination using multiple data modalities. Deep learning methods have shown promise in utilizing these multimodal data sources, outperforming single-modal models. However, integrating these heterogeneous data sources poses significant challenges in clinical decision-making. This study proposes an optimized multimodal CNN for a stacked ensemble model (OMCNNSE) for breast cancer prognosis. Our novel method involves the integration of the Tug of War (TWO) algorithm to optimize the hyperparameters of a convolutional neural network (CNN), enhancing feature extraction from three distinct multimodal datasets: clinical profile data, copy number alteration (CNA), and gene expression data. Specifically, we employ the TWO algorithm to optimize separate CNN models for each dataset, identifying optimal values for the hyperparameters. We then trained the three baseline CNN models using the optimized values through 10-fold crossvalidation. Finally, we utilize an ensemble learning approach to integrate the models’ predictions and apply an SVM classifier for the final prediction. To evaluate the proposed method, we conducted experiments on the METABRIC breast cancer dataset comprising diverse patient profiles. Our results demonstrated the effectiveness of the OMCNNSE approach for predicting breast cancer prognosis. The model achieved high AUC, accuracy, sensitivity, precision, and MCC, outperforming traditional single-modal models and other state-of-the-art methods.

A Statistical Model for Analyzing the Impact of GDP Components on the Manufacturing Sector in Saudi Arabia

Bader S Alanazi — 2024-12-19

This study examines the relationship between the Gross Domestic Product (GDP) of the manufacturing sector and various independent sectors in Saudi Arabia, utilizing a comprehensive dataset from 2010 to 2023 sourced from the General Authority for Statistics. A robust statistical model was developed to analyze these relationships, revealing significant insights into how changes in different sectors influence manufacturing GDP. Our findings indicate that increases in GDP from specific sectors, such as mining and construction, lead to growth in manufacturing GDP, while increases in other sectors, such as transport and community services, may negatively impact manufacturing performance. These insights are vital for policymakers, emphasizing the interconnectedness of economic sectors and the need for coordinated strategies. This research not only contributes to a deeper understanding of the dynamics shaping Saudi Arabia’s economic landscape but also provides a valuable foundation for future studies and policy interventions aimed at strengthening the manufacturing sector and promoting sustainable economic development.

Unraveling the Intuitionistic Octagonal Fuzzy Travelling Salesman Problem via Dhouib-Matrix-TSP1 Heuristic

Souhail Dhouib — 2024-12-18

The Travelling Salesman Problem (TSP) is an NP-hard problem of optimization that its goal is to obtain the shortest cycle among all cities that should be visited only once by a salesperson. The main goal of a salesperson is to visit each city only once and to obtain the distance traveled as well as the travelling costs as low as possible. In real-life and due to the absence of information, variables coming from experts’ collected data are usually uncertain and imprecise. In such cases, the decision maker cannot exactly expect the TSP cost. That’s why in this paper, the TSP under the intuitionistic octagonal fuzzy environment is considered and solved by adapting the very recent greedy method namely Dhouib-Matrix-TSP1 (DM-TSP1). This heuristic is very simple and it is composed of four steps. DM-TSP1 uses the Sum metric and is enriched with a ranking function. This current research work presents the first resolution of the TSP under the intuitionistic octagonal fuzzy domain. For this reason, new case studies are generated in order to carry out the experimental results. Moreover, a step-by-step execution of DM-TSP1 is detailed in order to prove its effectiveness.

Machine learning methods for modelling and predicting dust storms in Iraq

Heyam Hayawi — 2024-12-18

Dust storms are a significant problem that impacts humans, the environment, and the economy in Iraq, especially in the Baghdad, Nineveh, and Basra provinces, which have the most vital and substantial urban agglomerations in Iraq. These areas are heavily affected by dust storms. The monthly dust storm data and factors based on temperature, surface pressure, wind speed, wind direction, humidity, and precipitation were sourced from the Iraqi Meteorological Organization and Seismology and NASA from January 1981 to December 2022.

In this study, we used the principal components intended to reduce the interrelated variables and capture the components that account for at least 80\% of the total variance in the data set. Various supervised machine learning algorithms created a model to analyze and predict the monthly frequency of dust storms in the three provinces until March 2027.

Our findings indicate that the Additive Regression model, employing the IBk lazy algorithm for the Basrah and Nineveh regions and the KStar lazy algorithm for Baghdad, outperformed other methods in terms of accuracy. The results suggest a reduction in the occurrence of dust storms in the three provinces, with this downward trend projected to persist over the next 40 months.

Log-ergodicity: A New Concept for Modeling Financial Markets

Kiarash Firouzi — 2024-12-13

Although financial models violate ergodicity in general, observing the ergodic behavior in the markets is not rare. Policymakers and market participants control the market behavior in critical and emergency states, which leads to some degree of ergodicity as their actions are intentional. In this paper, we define a parametric operator that acts on the space of positive stochastic processes, transforming a class of positive stochastic processes into mean-ergodic processes. With this mechanism, we extract the data regarding the ergodic behavior hidden in the financial model, apply it to mathematical finance, and establish a novel method for pricing contingent claims. We provide some empirical examples and compare the results with existing ones to demonstrate the efficacy of this new approach.

Developing a Semi-parametric Zero-Inflated Beta Regression Model Using P-splines: Simulation and Application

Muhammad M. Seliem — 2024-11-25

Analyzing proportional data with excessive zeros and complex relationships presents a significant challenge in various fields. To address this, we propose a developing semiparametric Zero-Inflated Beta Regression (ZIBE) model incorporating P-splines. Our ZIBE.pb model offers a unique combination of flexibility and interpretability, allowing for the modeling of non-linear relationships and the identification of factors contributing to zero inflation. Extensive simulations demonstrate the ZIBE.pb model's superior model fit and predictive accuracy compared to existing parametric models. The proposed estimators exhibit minimum GD, AIC, BIC, and MSE, as confirmed by Monte Carlo simulation studies and real-world applications. Our ZIBE.pb model has broad applications in various fields, including political science, economics, and social sciences. To demonstrate its utility, we applied it to the Varieties of Democracy (V-Dem) dataset. In conclusion, our ZIBE.pb model offers a robust and versatile tool for analyzing proportional data with excessive zeros and complex relationships. Its ability to capture both linear and nonlinear effects, coupled with its interpretability, makes it a valuable asset for researchers across various domains.

A New Version of the Inverse Weibull Model with Properties, Applications and Different Methods of Estimation

Mohamed Ibrahim — 2025-02-08

A new extension of the inverse Weibull model is introduced and studied. Some of its statistical properties are derived. Di¤erent estimation methods are used for estimating the unknown parameter. We assessed the performance of all methods via simulation study. Two real data applications are used for comparing competitive estimation methods. The importance of the new model is demonstrated via two real data applications. The new model is much better than other competitive models in modeling the two real data sets.

A Weighted Exponentiated class of Distributions:Properties with Applications for Modelling Reliability Data

Gorgees Shaheed — 2025-01-19

lIn this study, we suggest the weighted Exp-G (WExp-G) continuous distributions as a novel class of continuous distributions with an additional shape parameter. Then we study the basic mathematical properties. We study Lindley and XGamma special cases. This model is exible for modelling right skew data sets. The hazard rate of this model is decraesing, increasing and bathtub shape. By performing a simulation analysis, we compared different common methods of estimation. Finally we analyzed and used lifetime, failure time and stress real data sets to illustrate the purposes. This model is perfrom better than other two-parameter distribution.

Fuzzified Clustering and Sample Reduction for Intelligent High Performance Distributed Classification of Heterogeneous Uncertain Big Data

Sherouk Samir Moawad — 2025-01-29

diverse datasets efficiently. This paper introduces a Fuzzified Clustering technique with sample reduction and distributed Parallel Classification (FCPC). Fuzzified clustering is particularly well-suited for Big Data as it enables the intelligent partitioning of datasets while managing uncertainties and overlapping data points. The FCPC technique takes advantage of this capability to reduce dataset size, capturing essential data structures and enhancing classification performance. Benchmark Big Data sets are used to compare FCPC with traditional classifiers, which require the entire dataset to fit in memory. Four classification techniques were evaluated in terms of classification evaluation metrics, namely, Accuracy, Area Under the ROC Curve, and F1 Score. The proposed model demonstrated improved classification predictive power with a sample reduction of approximately 90%, leading to enhanced performance and potential reductions in computational resources.

Implementing Automatic Microservices Detection in Business Processes Using Association Rules

Mohamed Daoud — 2024-12-12

Microservice-oriented architectures are increasingly becoming the preferred architectural style over monolithic systems, both in academic research and industrial applications. This shift is largely due to microservices' ability to deconstruct large, monolithic applications into smaller, independent, highly cohesive, and loosely connected services.However, the process of identifying appropriate microservices is a significant challenge, which, if not addressed adequately, could hinder the effectiveness and benefits of transitioning to this architectural style. In this paper, we introduce an innovative method based on association rules to automate the identification of microservices within a business process. This technique leverages association analysis to uncover latent correlations among the attributes of various activities. Activities sharing similar attributes are then grouped into the same microservice categories. To validate and demonstrate the practicality of our method, we conduct a case study focusing on a bicycle rental system.

A New Method in Machine Learning Adapted for Credit Risk Prediction of Bank Loans

Zaynab Hjouji — 2024-12-12

The recent global financial crisis has significantly impacted the financial system, leading to major bank failures and prompting a reevaluation of credit risk management models. Given its critical role in maintaining banking stability, effective credit risk forecasting methods are essential. In light of this, various studies have introduced techniques to analyze, detect, and prevent bank credit defaults. In this paper, we present a new approach for predicting credit risk, known as the “Method of Separating the Learning Set into Two Balls.” This method involves partitioning a learning set into two distinct categories: the "Performing Ball," which contains feature vectors of customers with non-defaulting credits, and the "Non-Performing Ball," which includes vectors of customers with defaulting credits. To predict a customer’s default risk, it is sufficient to determine which ball their feature vectors belong to. If a customer’s vectors do not fall into either category, additional analysis is required for making a credit decision. We evaluated the performance of this method through extensive experimental tests and a comparative analysis. The findings suggest that our approach shows considerable promise for enhancing credit risk prediction in the banking sector.

Developing a Collaborative Learning Environment For ITS: A New Model Based on IMS-Learning Design

Lahmadi Youssef — 2024-12-05

Intelligent Tutoring Systems (ITS) represent a significant advancement in educational technology, evolving from computer-assisted teaching to more adaptive and interactive learning environments. This paper aims to delve into the methodological aspects of collaborative ITS, with a particular focus on the integration of IMS Learning Design (IMS-LD). The objective is to explore how IMS-LD is instrumental in designing and managing ITS and addressing the challenges of modern education systems. The proposed model addresses these challenges through three foundational principles: categorizing learning activities, clearly defining roles, and designating specific spaces for diverse activities. By integrating the learner model and IMS-LD, the model aims to enhance personalization and effectiveness, creating a more efficient, learner-centric system. The paper also discusses the development of meta-models for collaborative ITS, their correspondence with IMS-LD, and the challenges and benefits of model transformation techniques. The findings highlight the potential of ITS in providing adaptive and personalized learning experiences, fostering effective communication and collaboration among participants, and enhancing the overall quality of education through innovative technological integration.

Integrating LSTM and LOF: A Comprehensive Approach to Anomaly Detection in Healthcare Data

Heba Mostafa Mohamed — 2024-12-20

The Internet of Medical Things (IoMT) has grown substantially, facilitating extensive time series data accumulation within healthcare environments. Detecting anomalies within IoMT time series data is critical for identifying potential health hazards and ensuring patient safety. This study investigates the efficacy of merging Long Short-Term Memory (LSTM) neural networks with the Local Outlier Factor (LOF) algorithm for anomaly detection in time series data. LSTM networks are adept at grasping extended dependencies in sequential data, while LOF represents a potent unsupervised outlier identification technique. We introduce a unique methodology that capitalizes on the strengths of both LSTM and LOF to heighten anomaly detection accuracy. The proposed technique undergoes assessment via experiments on actual IoMT datasets including WUSTL-EHMS and Thyroid\_Diff, demonstrating consistent performance across diverse healthcare scenarios. A real-time simulation was conducted to assess the feasibility of deploying the framework in practical IoMT environments. Through comprehensive experimentation and analysis, we show its effectiveness. The results of the proposed method reveal a promising ability to precisely pinpoint anomalies, offering a valuable resource to healthcare professionals to quickly identify irregular patient conditions.

A Sine-Cosine Algorithm Approach for Optimal PV and D-STATCOM Integration in Distribution Systems

Oscar Danilo Montoya — 2024-11-29

This study proposes the use of the Sine-Cosine Algorithm (SCA) to optimize the integration of solar photovoltaic (PV) systems and distribution static compensators (D-STATCOMs) in medium-voltage distribution networks. The SCA is applied to determine the optimal placement and sizing of both PV units and D-STATCOMs, utilizing a hybrid discrete-continuous approach for codifying the solutions. A power flow analysis, based on the successive approximations method, is employed to assess system performance, including voltage regulation and power distribution. The optimization is carried out using a master-slave framework, where the SCA handles the optimization process, and the power flow model evaluates the technical outcomes. Case studies on 33- and 69-bus systems reveal that the SCA achieves significant reductions in system losses, with improvements of approximately $35.5227$\% and $35.6331$\%, respectively. Moreover, the SCA demonstrates computational efficiency, outperforming other methods such as the Vortex Search Algorithm (VSA) and previous benchmarks. All simulations and validations were conducted using MATLAB 2024a, confirming the SCA’s robustness for this application.

Central Metric Dimension of Rooted Product Graph

Liliek Susilowati — 2024-10-18

The Central metric dimension is a type of metric dimension on graph. Some special graphs for which the central metric dimension have been found include path graph, cycle graph, complete graph, and complete bipartite graph. The aim of this study is to determine the central metric dimension of rooted product graph. Let G be a connected graph of order n and H is a sequence of n rooted graphs H₁, H₂, H₃, ..., H_n . The rooted product graph G and H denoted by GoH . In this paper, we determine the central metric dimension of rooted product graph which denoted by dim_cen( GoH)_. The results obtained for GoH where H is a sequence of rooted graphs that all have the same radius and the rooted vertex is the central vertex. For H is a sequence of rooted cycle graph, the cycle with the largest radius has an impact on the central set, while the central metric dimension is affected by the central set of GoH . For H is a sequence of rooted complete graph, the central set is affected by the central set of a graph G, while the central metric dimension is affected by the central set of the graph G.

On The Packing k-Coloring of Some Family Trees

Arika Indah Kristiana — 2024-09-09

All graphs in this paper are simple and connected. Let $G=(V,E)$ be a graph where $V(G)$ is nonempty of vertex set of $G$ and $E(G)$ is possibly empty set of unordered pairs of elements of $V(G)$. The distance from $u$ to $v$ in $G$ is the length of a shortest path joining them, denoted by $d(u,v)$. For some positive integer $k$, a function $ c:V(G)\rightarrow \{1,2,...k\} $ is called packing $k-$coloring if any two not adjacent vertices $u$ and $v$, $c(u)=c(v)=i$ and $d(u,v)\geq i+1$. The minimum number $k$ such that the graph $G$ has a packing $k-$coloring is called the packing chromatic number, denoted by $\chi_\rho(G) $. In this paper, we investigate the packing chromatic number of some family trees, namely centipede, firecracker, broom, double star and banana tree graphs.

Moments and Inferences based on Generalized Order Statistics from Benktander Type II Distribution

Zaki Anwar — 2025-01-29

In this paper, we employ generalized order statistics to investigate the moment properties of the Benktander Type II distribution. Through this approach, we derive precise and explicit formulas for single moments and establish recurrence relations for single and product moments. Additionally, we present a characterization of the Benktander Type II distribution, accompanied by further implications regarding moments of record values and ordinary order statistics. We estimate the unknown parameters of the Benktander Type II distribution using Maximum Likelihood (ML) estimation for generalized order statistics (gos). Subsequently, we conduct simulation studies encompassing order statistics. The efficacy of the obtained ML estimates is evaluated through comprehensive simulation analyses, focusing on various moments and their relative mean squared errors. This research contributes to understanding the Benktander Type II distribution's properties and provides valuable insights into its parameter estimation using generalized order statistics.

The Odd Generalized Rayleigh Reciprocal Weibull Family of Distributions with Applications

Regent Retrospect Musekwa — 2025-02-01

We introduced a novel family of models in this paper, which we named the odd-generalized Rayleigh reciprocal
Weibull-G (OGR-RW-G) family. This family is noteworthy because it applies the T-X model construction technique to the generalized Rayleigh reciprocal Weibull model, addressing the inflexibility limits associated with traditional models and allowing one to use any baseline distribution. We examine some valuable statistical inferences from the OGR-RW-G, including its probability density function (pdf) represented in a linear fashion, its order statistics’ pdf, moments, residual life functions and R´enyi entropy. Additionally, the hazard rate functions (hrfs) and pdfs of a few particular models are determined to have analytical shapes. The OGR-RW-G model parameters are determined by the widely recognized maximum likelihood estimation (MLE) technique. We also perform a simulation exercise to evaluate the performance of the MLEs. Ultimately, the utility of the OGR-RW-G family is demonstrated by using the odd generalized Rayleigh reciprocal Weibull Burr-XII (OGR-RW-BXII) example of the OGR-RW-G to three distinct datasets. In actuality, the four parameter OGR-RW-BXII outperforms the four parameter non-nested models and some nested models that are presented.

Application of the Periodic Self-Exciting Threshold Autoregressive Model

Nesrine Bezziche — 2025-01-29

In this paper, we analyze Algerian temperature data using the periodic self-exciting threshold autoregressive (PSETAR) model. Despite the significant advantages offered by the periodic SETAR model in capturing seasonal and threshold-based behaviors, it remains underutilized in practical applications. The goal of this work is to demonstrate the utility of this model by applying it to Algeria's temperature series. We examine the properties of the model and discuss its estimation using the least squares method. The linearity is tested using the likelihood ratio test, and we extend the local asymptotic normality to p regimes. This analysis provides a deeper understanding of the temperature dynamics in Algeria, highlighting the model's ability to capture seasonal variations and thresholds.