Journal of Artificial Intelligence and Metaheuristics

This review examines the role of feature selection in data analytics and its growing use in data-intensive fields, including cybersecurity, healthcare, and finance. It surveys foundational concepts and current applications, with particular attention to reducing dimensionality and improving the interpretation of complex relationships in healthcare data. The review covers optimization-based methods, swarm intelligence, and feature-selection approaches for heart-disease diagnosis, intrusion detection, and other classification tasks. It also discusses hybrid Ant Colony Optimization and Grey Wolf Optimizer methods, alongside the Improved Salp Swarm Algorithm (ISSA). The reviewed studies illustrate that feature selection is more than a preprocessing step: it is an important component of machine learning and data science that can improve model efficiency, interpretability, and classification performance. Challenges and open research directions are summarized to support future work.

Digital transformation is steadily changing states and organizations, making them more competitive while creating opportunities for economic growth and prosperity. It enables expanded educational opportunities, universal Internet access, and a supportive environment for development. This paper highlights the role of the Communication and Information Technology industry in achieving sustainable development. It reviews the 17 Sustainable Development Goals (SDGs) and explains how information technology affects each goal. It also tracks progress toward the SDGs and existing e-governance and big-data initiatives in four countries, before examining digital transformation in Egypt and its contribution to sustainable development.

Student-centered analysis of academic performance is also the most important aspect in improving education by being able to determine what measures work best, individualized learning approaches, and intervention programs. In this study, we performed a detailed analysis based on the “Students Performance in Exams” dataset and different regression methods to estimate students’ grades. We sought to assess the functioning of numerous metrics and determine an optimal model for this task. Our descriptive analysis identified meaningful trends within this dataset, as it includes central factors like gender, race/ethnic diversity-based status of a student, and parental education level based on which the children are catered to by informing them about important lunches and test preparation courses alongside scores in “Math,” “Readings,” “Writing,” etc. We used a wide range of regression models: XGBoost, CatBoost, GradientBoostingRegressor, etc. Metrics like Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and Average Marginal Loss were used to assess each model rigorously. Importantly, the XGBoost model gave out an MSE value of 0.028, which was the best among all values obtained from various other models. The superiority of the XGBoost model is supported by the excellent performance that was reported across many metrics. This work can be important for informing educational practitioners and policymakers regarding the best possible accurate and realistic model that would predict the students’ outcome results. Educational data analytics incorporating the XGBoost model can be used for the customization of interventions and mapping resource allocation while promoting a results-oriented approach based on data in education. This study is a step towards the accumulation of knowledge on educational data analytics. It can serve as a background for further research aimed at improving predictive models regarding student performance.

This paper deals with a pivotal part of educational data analytics, aiming to increase the accuracy and interpretability of student performance prediction models. The cornerstone of our method is the innovative application of binary Waterwheel Plant Algorithm (bWWPA) in feature selection. As we can see, an essential part of any model is the predicted values, which correctly define all the characteristics of this model. Practically, we begin with solid data pre-processing, which incorporates data cleaning and missing values, duplicate removal, and data transformation in order to get model input as optimally as possible. Preceding the application of bWWPA, we employ an ensemble of regression machine-learning models. We set up a baseline for predictive capability, obtaining initial outcomes with an average Mean Squared Error (MSE) of 0.064. The following feature-selection phase shows the algorithm’s ability to recognize important elements and, as a result, improve model effectiveness and explanatory power. The comparative analyses after feature selection point to refined gains in the model, with the refined models reporting a lower MSE of 0.032. These findings, methodologically, add to student performance prediction. Accordingly, they emphasize the decisive status of feature selection in improving models. The paper’s significance extends to teachers, institutions, and researchers, giving insights into more precise and relevant student-success interventions.

Advanced methods are needed for fast and reliable detection of cardiovascular illnesses, which continue to be a primary source of morbidity and death globally. Using deep learning, this research presents a new method, dubbed “DeepLearnCardia,” for analyzing electrophysiological data in cardiac bioengineering. To improve the analysis of cardiac electrophysiological data and provide a complete solution for arrhythmia prediction, the proposed technique combines wavelet transformations, attention processes, and multimodal fusion. Data preprocessing, feature extraction using wavelets, temporal encoding using Long Short-Term Memory (LSTM) networks, an attention mechanism, multimodal fusion, and spatial analysis with Convolutional Neural Networks (CNNs) are all components of this technique. In order to train the model, we use an adaptive optimizer and binary cross entropy as the loss function. Key performance metrics such as accuracy, sensitivity, specificity, precision, F1 score, and area under the ROC curve (AUC-ROC) are used to compare the proposed method’s performance to that of six established methods: Signal Pro Analyzer, Electro Cardio Suite, Bio Signal Master, Cardio Wave Analyzer, EKG Precision Pro, and Heart Stat Analyzer. The results suggest that the proposed technique is superior to the state-of-the-art in cardiac signal analysis across all criteria. The suggested technique not only requires less resources, but also trains and infers more quickly and uses less of them.