Hossam El-Din Moustafa

Accurate solar radiation forecasting is essential for improving the reliability of photovoltaic energy generation and supporting effective solar battery management, particularly because solar radiation is highly variable and depends on nonlinear interactions among meteorological and temporal factors. Although conventional prediction methods can provide useful estimates, they often struggle to capture the sequential behavior of solar radiation caused by daily sunlight cycles, atmospheric variation, and changing weather conditions. Therefore, this study aims to develop and evaluate deep learning models for predicting Solar_radiation using meteorological data collected from the HI-SEAS weather station over four months, from September to December 2016, where the main input variables include temperature, humidity, pressure, wind direction, wind speed, and time-related features. Five recurrent deep learning models were implemented and compared, namely Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), Bidirectional Long Short Term Memory (BiLSTM), and Attention-LSTM. Before model training, the dataset was preprocessed by handling missing values, checking temporal consistency, arranging the observations chronologically, and applying Min–Max normalization to ensure stable learning. Model performance was assessed using multiple regression metrics, including MSE, RMSE, MAE, MBE, correlation coefficient, R2, RRMSE, NSE, and WI. The experimental results showed that BiLSTM achieved the best overall forecasting performance, with an MSE of 0.0014, RMSE of 0.0379, MAE of 0.0182, MBE of 0.0039, correlation coefficient of 0.9750, R2 of 0.9494, RRMSE of 0.3645, NSE of 0.9494, and WI of 0.9865. GRU and RNN also produced competitive results, achieving RMSE values of 0.0381 and 0.0382 and R2 values of 0.9489 and 0.9486, respectively, while Attention-LSTM showed comparatively lower performance with an RMSE of 0.0492 and R2 of 0.9149. These findings indicate that recurrent deep learning models are effective for learning nonlinear and temporal patterns in solar radiation data, with BiLSTM providing the most accurate and reliable predictions. The proposed forecasting framework can support photovoltaic energy planning and solar battery decision-making by estimating future solar radiation levels and helping determine whether solar energy utilization will be feasible under expected weather conditions.

American scientific publishing group

Diabetes mellitus (DM) is a chronic metabolic disorder that poses significant health risks and global economic burdens. Early prediction and accurate diagnosis are crucial for effective management and treatment. This study presents an ensemble machine learning-based model designed to predict and diagnose Diabetes Mellitus using clinical and demographic data. The proposed approach integrates multiple machine learning algorithms, including random forest (RF), extreme gradient boosting (XGB), and logistic regression (LR), to leverage their individual strengths and enhance the entire performance. The ensemble model was trained and validated on multiple comprehensive datasets. Performance measures demonstrate the robustness of proposed model and its reliability in distinguishing diabetic cases from non-diabetic cases after applying several preprocessing steps. This work ensures the capability of machine learning in advancing healthcare by providing efficient, data-driven tools for diabetes management, aiding clinicians in early diagnosis, and contributing to personalized treatment strategies. Comparative analysis against standalone models highlights the superior predictive capabilities of the ensemble approach. Results had shown that ensemble model achieved an accuracy of 96.88% and precision of 89.85% outperforming individual classifiers.

This paper introduces OLIVE-CAD, a novel Computer-Aided Diagnostics system designed for the real life, on-site detection of olive leaf diseases. The core of the system is a YOLOv12-based convolutional neural network model, which was trained on a comprehensive dataset of 11,315 olive leaf images. The images were categorized into 'Aculus', 'Scab', and 'Healthy,' with the dataset divided for training (70%), evaluation (20%), and real-world testing (10%). The key contribution of this work is the end-to-end integration of a custom, field-deployable Computer-Aided Diagnostics system. The trained YOLOv12 model achieved a mean average precision of 98.2% and mean average recall of 95.4%, while the model achieves class-specific evaluation precision of 95.3% and recall of 97.7% for 'Healthy' class; 97.9% precision and 88.3% of recall for 'Aculus' class; and precision of 94.3% and 95.4% of recall for 'Scab' class. OLIVE-CAD enables the storage of the immediate disease diagnostic outcomes to a predesigned database, providing a practical, deployable solution for agricultural applications. The research recommends an IoT-Based real-time central operation diagnostic monitoring system as future work.

Microsatellite instability (MSI) is considered a significant biomarker for gastrointestinal (GI) cancer prognosis and treatment planning. Traditionally, molecular assays such as polymerase chain reaction (PCR) testing and immunohistochemistry (IHC) have been used to determine MSI status. Despite their effectiveness, these methods are labor-intensive and time-consuming .MSI tumors, in particular, are known to respond better to immunotherapy due to their high mutational burden and increased immunogenicity, making accurate MSI assessment vital for selecting appropriate treatments. The ability to classify MSI status directly from histopathological images offers a faster and less invasive alternative that could greatly enhance clinical decision-making and help guide personalized immunotherapy strategies. In this study, an average ensemble classifier combining two pretrained convolutional neural network (CNN) architectures—InceptionResNet-v2 and Xception—has been developed to improve the robustness and accuracy of MSI prediction from histological images. The suggested ensemble model outperformed earlier methods with an accuracy of 96.97% and an area under the curve (AUC) of 99.57%. These results demonstrate that ensemble-based deep learning models can accurately and reliably classify MSI from histological slides, facilitating more personalized treatment decisions and ultimately improving outcomes for patients receiving immunotherapy.

Asthma, a prevalent and complex chronic respiratory disease, imposes a growing burden on global healthcare systems. This survey investigates the evolving role of Artificial Intelligence (AI), particularly Machine Learning (ML) and Deep Learning (DL), in enhancing early detection, personalized care, and real-time monitoring of asthma. Drawing on a structured synthetic dataset encompassing demographic, environmental, lifestyle, and clinical variables, the study applies rigorous data preprocessing and comparative model evaluation across multiple ML algorithms, including Logistic Regression, Support Vector Machine, Random Forest, Gradient Boosting, and XGBoost. Among these, the XGBoost classifier outperforms others, achieving 98% accuracy and an AUC of 0.94, demonstrating its robustness for structured health data. Additionally, the integration of wearable and real-time sensor data is identified as a critical future direction to further improve predictive performance and clinical applicability. This review highlights the potential of AI-driven approaches to revolutionize asthma care by enabling timely interventions, reducing hospitalizations, and supporting individualized management strategies.

Stroke is among the leading causes of death, especially among old adults. Thus, the mortality rate and severe cerebral disability can be avoided when stroke is diagnosed at its early stages, followed by subsequent treatment. There is no doubt that healthcare specialists can find the necessary solutions more effectively and instantly with the help of artificial intelligence (AI) and machine learning (ML). In this study, we used ML classifiers and explainable artificial intelligence (XAI) to predict stroke. Six different ML classifiers that trained on available datasets for stroke patients. Six feature selection methodologies were used to extract essential features from the dataset. The XAI methods applied (Shapley Additive Values (SHAP), ELI5, and Local Interpretable Model-agnostic Explanations (LIME)). This study will help medical practitioners to effectively manage their patients and develop a screening tool that incorporates technologies to revolutionize the approach to preventing and treating stroke.

It is no doubt that diabetes is considered one of the most common chronic diseases. Diabetes patients have high risk of diseases like renal failure, heart stroke, nerve and eye damage that can lead to blindness. Detection and prediction of diabetes mellitus is not a very easy process. Nevertheless, the cost of tests is high. Hospitals being busy therefore it could be a revolutionary if one could know the risk of being diabetic with no need to visit hospitals. This could only be done through artificial intelligence. In this paper, a classification model was proposed for diabetes mellitus classification and pre-diction, so that early diagnosis as well as treatment could prolong patients’ lives and minimize risk factors. The classification of datasets in medical healthcare is hindered by the problem of having suitable datasets. Proper processing was performed through null values imputation, normalization and encoding. Supervised algorithms were applied to ensure the effectiveness of the proposed model such as Random Forest (RF), Extreme Gradient Boosting (XGB) and Neural Network (NN). Results were compared using five performance metrics; accuracy, precision, f1-score, recall and run time. Training and testing are performed on two datasets. Results demonstrated that RF has overtaken both remaining techniques by achieving 80.5% accuracy compared to 79.65% for XGB and 76.36% for NN on the first dataset. While the second dataset results indicated RF superiority among remaining models by achieving an accuracy of 97.11% compared to 93.38% and 93.26 for NN and XGB respectively.

We extend our heartfelt gratitude to our diligent and meticulous reviewers of Mansoura Engineering Journal for their unwavering commitment and invaluable contributions throughout the year 2024. Their dedication to maintaining the highest standards of academic excellence has been instrumental in shaping the scholarly discourse within the engineering community. The editorial and review process is the cornerstone of academic publishing, ensuring the integrity, relevance and rigor of the research disseminated to the global scientific community. The tireless efforts of the editors in overseeing the publication process with precision and efficiency, coupled with the meticulous scrutiny of