A Comprehensive Review on Optimizing Machine Learning Models for Early Detection and Forecasting of Monkeypox Outbreaks

Ahmed El-Sayed; Ahmed M.; Sekar Kidambi; Marwa M.

doi:https://doi.org/10.54216/JAIM.080102

Fusion: Practice and Applications | 2025

A Comprehensive Survey on AlexNet improvements and fusion techniques

Machine- and deep-learning techniques have been used in numerous real-world applications. One of the famous deep-learning methodologies is the Deep Convolutional Neural Network. AlexNet is a well-known global deep convolutional neural network architecture. AlexNet significantly contributes to solving different classification problems in different applications based on deep learning. Therefore, it is necessary to continuously improve the model to enhance its performance. This survey study formally defined the AlexNet architecture, presented information on current improvement solutions, and reviewed applications based on AlexNet improvements. This work also presents a simple survey based on a fusion of AlexNet with different machine-learning techniques for recent research in biomedical applications. In the survey results for about 11 research papers for both improvement and fusion techniques of AlexNet, it was clear that the fusion was the superior one with 99.72, and the improved one was 99.7%. In the conclusion and discussion section, there was a comparison between the improved techniques and fusion techniques of AlexNet and a proposal for future work on AlexNet development.

groups

Bahaa S. Rabi mail -

Ayman S. Selmy mail -

Wael A. Mohamed mail

link https://doi.org/10.54216/FPA.170210

Volume & Issue

Vol. Volume 17 / Iss. Issue 2

Details open_in_new

Fusion: Practice and Applications | 2025

Enhancing Object Detection and Classification Using White Shark Optimization with Deep Learning on Remote Sensing Images

Remote sensing (RS) object detection is extensively applied in the fields of civilian and military. The important role of remote sensing is to identify objects like planes, ships, harbours airports, etc., and then it can attain position information and object classification. It is of considerable importance to use RS images for observing the densely organized and directional objects namely ships and cars parked in harbours and parking areas. The object detection (OD) process involves object localization and classification. Due to its wide coverage and longer shooting distance, Remote sensing images (RSIs) have hundreds of smaller objects and dense scenes. Deep learning (DL), in particular convolution neural network (CNN), has revolutionized OD in different fields. CNN is devised to automatically learn the hierarchical representation of data, which makes them fit for feature extraction. Hence, the study proposes a new white shark optimizer with DL-based object detection and classification on RSI (WSODL-ODCRSI) method. The purpose of the WSODL-ODCRSI model is to classify and detect the presence of the objects in the RSI. To accomplish this, the WSODL-ODCRSI model uses a modified single-shot multi-box detector (MSSD) for the OD process. The next stage of OD is the object classification process, which takes place with the use of the Elman Neural Network (ENN) algorithm. The WSO algorithm is exploited as a parameter-tuning model for improving the object classification results of the ENN approach. The stimulated study of the WSODL-ODCRSI algorithm has been established on the benchmark data set and the outcomes underlined the promising performance of the WSODL-ODCRSI model on the object process of classification

groups

Reda Salama mail

link https://doi.org/10.54216/FPA.170211

Volume & Issue

Vol. Volume 17 / Iss. Issue 2

Details open_in_new

Fusion: Practice and Applications | 2025

Analysis of Objective Functions for Ribonucleic Acid Multiple Sequence Alignment Fusion Based on Harmony Search Algorithm

Four kinds of smaller molecules known as ribonucleotide bases-adenine (A), cytosine (C), guanine (G), and uracil (U) combine to form the linear molecule known as ribonucleic acid (RNA). Aligning multiple sequences is a fundamental task in bioinformatics. This paper studies the correlation of different objective functions applying to RNA multiple sequence alignment (MSA) fusion generated by the Harmony search-based method. Experiments are performed on the BRAliBase dataset containing different numbers of test groups. The correlation of the alignment score and the quality obtained is compared against coffee, sum-of-pairs (SP), weight sum-of-pairs (WSP), NorMD, and MstatX. The results indicate that COFFEE and SP objective functions achieved a correlation coefficient (R²) of 0.96 and 0.92, respectively, when compared to the reference alignments, demonstrating their effectiveness in producing high-quality alignments. In addition, the sum-of-pairs takes less time than the COFFEE objective function for the same number of iterations on the same RNA benchmark.

groups

Mubarak Saif mail -

Rosni Abdullah mail -

Mohd. Adib Hj. Omar mail -

Abdulghani Ali Ahmed mail -

Nurul Aswa Omar mail -

Salama A. Mostafa mail

link https://doi.org/10.54216/FPA.170201

Volume & Issue

Vol. Volume 17 / Iss. Issue 2

Details open_in_new

Fusion: Practice and Applications | 2025

The Detection of Glaucoma in Fundus Images Based on Convolutional Neural Network

Glaucoma is a common disease affecting the human retina, primarily caused by elevated intraocular pressure. Early intervention is crucial to prevent damage to the affected organs, which could lead to their dysfunction. This paper focuses on enhance diagnosis accuracy of the system to determine if a patient is at risk of developing glaucoma. In this paper a novel convolutional neural network (CNN) designed, specifically for the detection of glaucoma in fundus images. This architecture optimizes for the unique characteristics of fundus imagery, enhancing detection accuracy, and also compiled a large and diverse dataset of fundus images, crucial for training and validating our CNN model. The dataset includes a significant number of images with detailed annotations, ensuring robust model training. In addition, implemented sophisticated image preprocessing methods to enhance the quality of the fundus images. These techniques, including noise reduction and contrast enhancement, significantly improve the input data quality for the CNN. The system operates in three stages. First, it preprocesses the image by cropping, enhancing, and resizing it to a consistent 256×256 pixels. Next, it employs an advanced feature extraction to analyses key features of the optic disc and optic cup in retinal images. Finally, the Soft-Max function classifies the images, identifying those with glaucoma and distinguishing them from normal eye samples. The model's performance was thoroughly evaluated using various metrics like accuracy, Sensitivity, specificity, and the area under the curve are metrics used to evaluate the performance of a diagnostic test. Sensitivity measures the test's ability to correctly identify positive cases, specificity assesses its accuracy in identifying negative cases, and the area under the curve indicates the overall effectiveness of the test across different thresholds. The results achieved by the proposed system were thoroughly analyzed, revealing a high accuracy rate in glaucoma classification, reaching 99%.

groups

Ali Yakoob Al-Sultan mail

link https://doi.org/10.54216/FPA.170202

Volume & Issue

Vol. Volume 17 / Iss. Issue 2

Details open_in_new

Fusion: Practice and Applications | 2025

Computer Aided Brain Tumor Diagnosis using Coati Optimization Algorithm with Explainable Artificial Intelligence Approach

Brain tumors (BT) are a difficult and dangerous medical condition, and the accurate and early analysis of these tumors is crucial for suitable treatment. Explainability in clinical image diagnosis role a vital play in the correct analysis and treatment of tumors that supports medical staff's optimum understanding of the image analysis performances rely upon deep methods. Artificial intelligence (AI), in certain deep neural networks (DNNs) has attained remarkable outcomes for clinical image analysis in many applications. However, the need for explainability of deep neural approaches has been assumed that major restriction before executing these approaches in medical practice. Explainable AI, or XAI, is a vital module in this context as it supports medical staff and patients in understanding the AI's decision-making model, enhancing trust and transparency. It leads to optimum patient care and performance but making sure that medical staff can make learned decisions depends on AI-driven insights. Therefore, this study develops a novel Computer-Aided Brain Tumor Diagnosis using Coati Optimization Algorithm with an Explainable Artificial Intelligence (CABTD-COAXAI) approach. The purpose of the CABTD-COAXAI technique is to exploit XAI and hyperparameter-tuned deep learning (DL) approaches for automated BT analysis. To accomplish this, the CABTD-COAXAI technique follows a Gaussian filtering (GF) based noise removal process. Besides, the CABTD-COAXAI technique utilizes the EfficientNetB7 methods for the feature extraction process. Additionally, the hyperparameter tuning of the EfficientNetB7 method is performed by the use of COA. Furthermore, the classification of the BT process can be performed by the usage of a convolutional autoencoder (CAE). Finally, the CABTD-COAXAI system combines the XAI method named LIME to effectively understand and explainability of the black-box model for automated BT diagnosis. The simulation result of the CABTD-COAXAI technique has been tested on a benchmark BT database. The extensive outcomes inferred that the CABTD-COAXAI method reaches superior performance over other models in terms of different measures

groups

Wajdi Alghamdi mail

link https://doi.org/10.54216/FPA.170203

Volume & Issue

Vol. Volume 17 / Iss. Issue 2

Details open_in_new

Fusion: Practice and Applications | 2025

Elevating Diagnostic Accuracy: Advanced GAN-Enhanced High-Resolution Medical Imaging for Superior Disease Detection

Advanced imaging in medical has become crucial in the early identify diseases because they reveal the important structural features of the human body. But it is almost impossible to get such high resolution images in real life situation due to the factors such as image capture and processing equipment, and environmental factors that affect the outcome of the image. This work proposes a sub-type of GAN that is used in enhancement of images particularly in medical fields. The generator of the Med-GAN extracts a high-resolution image from a low-resolution one with the help of novel features learned by the model. The approach of reconstructing high resolution from multiple parallel streams of lower resolution employs deconvolution algorithms with multiple scale fusions that produce better high resolution representations as compared to the technique of bilinear interpolation. The performances of the proposed Med-GAN are tested on two publicly available COVID-19 CT datasets and one private medical image dataset which shows that the proposed method outperforms the existing methods in performance comparisons. Consequently, for PSNR, the score improves from 24.103 dB corresponding to the Initial Approach of the “BRaTS (FLAIR)” dataset to 25.496 dB for the Proposed Method; whereas for SSIM the score increases from 0.782 to 0.812.se types of high-resolution images are usually impossible to get due to limits in imaging devices, environmental conditions, and human factors. This work proposes the Med-GAN: an Enhanced Super-Resolution Generative Adversarial Network tuned for medical image enhancement. The Med-GAN generator learns high-resolution representations from low-resolution images via advanced feature extraction methods. Deconvolution algorithms with multi-scale fusions recover better high-resolution representations from multiple parallel streams of lower resolutions in this approach compared to traditional bilinear interpolation methods. Evaluated on two publicly available COVID-19 CT datasets and one custom medical image dataset, the proposed Med-GAN significantly outperforms existing techniques in performance comparisons. In particular, PSNR rises from 24.103 dB for the "BRaTS (FLAIR)" dataset in the Initial Approach to 25.496 dB in the Proposed Method, while SSIM increases from 0.782 to 0.812. If that is the case then it could be said that the solution of the proposed Med-GAN is one of the most realistic means for improving the quality of medical images and therefore contributes to better diagnostics of diseases

groups

Vathana D. mail -

Babu S. mail

link https://doi.org/10.54216/FPA.170214

Volume & Issue

Vol. Volume 17 / Iss. Issue 2

Details open_in_new

Fusion: Practice and Applications | 2025

A Comparative Analysis of Feature Extraction Techniques for Fake Reviews Detection

The current Internet era is characterized by the widespread circulation of ideas and viewpoints among users across many social media platforms, such as microblogging sites, personal blogs, and reviews. Detecting fake reviews has become a widespread problem on digital platforms, posing a major challenge for both consumers and businesses. Due to the ever-increasing number of online reviews, it is no longer possible to manually identify fraudulent reviews. Artificial intelligence (AI) is essential in addressing the problem of identifying fake reviews. Feature extraction is a crucial stage in detecting fake reviews, and successful feature engineering techniques can significantly improve the accuracy of opinion extraction. The paper compares five feature extraction methods for multiple opinion classification using Twitter on airline and Borderland game reviews. FastText with X-GBoost classifier outperformed all other techniques, achieving 94.10% accuracy on the airline dataset and 100% accuracy in Borderland game reviews.

groups

Zahraa Fadhel mail -

Hussien Attia mail -

Yossra Hussain Ali mail

link https://doi.org/10.54216/FPA.170212

Volume & Issue

Vol. Volume 17 / Iss. Issue 2

Details open_in_new

Fusion: Practice and Applications | 2025

Advancing Early Cardiovascular Disease Prediction Model using Improved Beluga Whale Optimization with Ensemble Learning via ECG Signal Analytics

Cardiovascular Disease (CVD) mainly affects the blood vessels and heart such as coronary artery disease, stroke, and heart failure. Early recognition is vital for on-time intervention and enhanced patient results. CVD is a major issue in society nowadays. When compared to the non-invasive model, the electrocardiogram (ECG) is the most effective approach for identifying cardiac defects. However, ECG analysis needs an experienced person with high knowledge and basically, it is a time-consuming task. Emerging a new technique to identify the disease at an early stage increases the quality and efficacy of medicinal care. A state-of-the-art technologies like machine learning (ML) and artificial intelligence (AI) have been gradually being used to increase the efficacy and accuracy of CVD recognition, permitting for faster and more exact analysis, and finally contributing to superior management and prevention tactics for CV health. This research paper designs an Early Cardiovascular Disease Prediction using an Improved Beluga Whale Optimizer with Ensemble Learning (ECVDP-IBWOEL) approach via ECG Signal Analytics. The main intention of the ECVDP-IBWOEL system is to forecast the presence of CVD at the early stage using EEG signals. In the ECVDP-IBWOEL method, the primary phase of data preprocessing is initially implemented to convert the input data into a well-suited layout. Also, the ECVDP-IBWOEL technique follows an ensemble learning (EL) process for CVD detection comprising three models namely long short-term memory (LSTM), deep belief networks (DBNs), and stacked autoencoder (SAE). Finally, the IBWO algorithm-based hyperparameter tuning process takes place which can boost the classifier results of the ensemble models. To certify the enhanced results of the ECVDP-IBWOEL system, an extensive experimental study is made. The experimentation outcomes stated that the ECVDP-IBWOEL system underlines promising performance in the CVD prediction process

groups

Hassan A. Alterazi mail

link https://doi.org/10.54216/FPA.170213

Volume & Issue

Vol. Volume 17 / Iss. Issue 2

Details open_in_new

Journal of Artificial Intelligence and Metaheuristics | 2024

An AI-Based System for Predicting Renewable Energy Power Output Using Advanced Optimization Algorithms

Accurate generation forecasting of Renewable Energy Sources (RES) is becoming more and more crucial for effective grid operation and energy management as RES are incorporated into the electrical grid. Because Machine Learning (ML) and Deep Learning (DL) algorithms can learn complicated relationships from data and provide accurate forecasts, they have become more popular than traditional forecasting approaches, which have limits.  This article examines the state of the art and future directions in the field of ML and DL-based forecasting of renewable energy generation. This paper reviews the several approaches and models that have been used to project renewable energy. It also highlights the challenges, such as managing the uncertainty and unpredictability of renewable energy output, data accessibility, and model interpret ability. To sum up, this study emphasizes how important it is to develop accurate and dependable renewable energy forecasting models to facilitate the future transition to sustainable energy sources and enable the integration of RES into the electrical grid.

groups

Mona Ahmed Yassen mail -

Mohamed Gamal Abdel-Fattah mail -

Islam Ismail mail -

EL-Sayed M. El Kenawy mail -

Hossam El-Din Moustafa mail

link https://doi.org/10.54216/JAIM.080101

Volume & Issue

Vol. Volume 8 / Iss. Issue 1

Details open_in_new

Journal of Artificial Intelligence and Metaheuristics | 2024

A Comprehensive Review on Optimizing Machine Learning Models for Early Detection and Forecasting of Monkeypox Outbreaks

This is a significant problem in diagnosing zoonotic opportunistic 'emerging' diseases like Monkeypox, which require not only better diagnostics but also efficient, effective, and affordable diagnostics. This paper considers the possibilities of machine learning (ML), deep learning (DL), and optimization algorithms for diagnosing and predicting Monkeypox. The presently employed strategies can be enhanced because clinical and imaging data can be harnessed to drive these technologies for early detection and subsequent containment activities. Generally, in a review, the authors offer information on how the diagnostic processes using ML and DL result in enhanced accuracy, specificity, and sensitivity of models, thus reducing design reliabilities. Furthermore, outbreak data is subjected to predictive modeling analysis to establish patterns useful in helping risk managers and policymakers prepare to manage future outbreaks. This system poses a new diagnostic model for Monkeypox and other zoonotic diseases by incorporating these complex computational tools into the present healthcare systems. This advancement not only strengthens the diagnostic arsenal of zoonotic diseases but also expands the possibilities for the interception and prevention of such diseases in the future at the world level.

groups

Ahmed El-Sayed Saqr mail -

Ahmed M. Elshewey mail -

Sekar Kidambi Raju mail -

Marwa M. Eid mail

link https://doi.org/10.54216/JAIM.080102

Volume & Issue

Vol. Volume 8 / Iss. Issue 1

Details open_in_new