International Journal of BIM and Engineering Science

The use of machine learning methods in healthcare has shown encouraging outcomes in terms of better patient care, more efficient use of resources, and streamlined operations. Traditional machine learning methods encounter difficulties when dealing with healthcare data due to its complexity and heterogeneity. Healthcare applications are a good fit for Gradient Boosting Machines (GBMs), which have become a formidable tool for structured data and predictive modelling jobs. Better healthcare system capabilities, including more precise forecasts and well-informed decisions, may be achieved by the integration of GBMs into a hybrid machine learning framework. Using GBMs and Reinforcement Learning (RL), the approach entails creating HealthCareAI, a Hybrid Fusion Learn-Enabled Software Product Line for Healthcare Optimization. Structured healthcare data, including patient information, medical records, and test results, are handled by GBMs. This includes data pre-processing, feature engineering, and GBM model training to forecast outcomes including illness diagnosis, treatment efficacy, and patient prognosis, among others. To optimize treatment planning and resource allocation, the HealthCareAI framework combines GBM models with CNNs for medical image processing and RL. The results show that GBMs in HealthCareAI are effective in boosting prediction accuracy and facilitating healthcare data-driven decision-making. A substantial improvement in predicting accuracy was shown across a range of healthcare jobs once Gradient Boosting Machines (GBMs) were included into HealthCareAI. When compared to more conventional machine learning approaches, GBM models improved illness prediction accuracy by an average of 15%. Even more significant improvements were seen in patient risk stratification, as GBMs successfully identified high-risk patients with an astounding sensitivity of 92% and specificity of 89%.

Efficient energy management in Wireless Sensor Networks (WSNs) is vital for extending network lifetime, particularly in applications requiring continuous monitoring in remote or challenging environments. This study proposes an energy-efficient multi-hop clustering approach for WSNs, utilizing intelligent swarm-based algorithms to optimize cluster formation and data routing. By applying Particle Swarm Optimization (PSO) and Ant Colony Optimization (ACO) techniques, the proposed method dynamically selects optimal cluster heads and minimizes energy consumption during multi-hop data transmission. The algorithm was evaluated on simulated WSN scenarios with varying node densities, achieving an average energy savings of 28% compared to traditional clustering methods and a 35% increase in network lifetime. Additionally, the proposed approach improved packet delivery ratio and reduced latency by 20% and 15%, respectively. This swarm-based, energy-efficient clustering framework is well-suited for applications in environmental monitoring, smart agriculture, and industrial automation, where prolonged network operation is essential.

This paper presents an adaptive optimization algorithm for personalized learning pathways in e-learning environments. The proposed algorithm dynamically adjusts the learning path for each student based on their performance, preferences, and learning behavior. By integrating machine learning techniques with a rule-based system, the algorithm optimizes content delivery and ensures a tailored learning experience that aligns with individual needs. The system continuously monitors learners’ progress, adapts to their evolving knowledge levels, and suggests the most relevant resources and activities to enhance engagement and comprehension. Experimental results demonstrate significant improvements in learning outcomes, reduced time to completion, and enhanced user satisfaction, making the approach a promising solution for personalized e-learning systems.

Underwater surveillance relies heavily on image quality, yet underwater environments present unique challenges, including low visibility, color distortion, and light scattering. This study proposes a bio-inspired image enhancement algorithm designed to address these challenges by mimicking adaptive mechanisms found in marine organisms. The algorithm integrates a multi-scale Retinex model with a bio-inspired filter based on visual properties of aquatic species, optimizing contrast and color balance for improved image clarity. Tested on various underwater image datasets, the proposed method achieved a 45% improvement in contrast enhancement and a 38% reduction in color distortion compared to traditional enhancement techniques. Furthermore, the Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM) improved by 42% and 35%, respectively. The results demonstrate the algorithm’s effectiveness in enhancing visibility and detail, enabling more accurate object detection and classification in underwater surveillance. The bio-inspired approach offers a practical solution for underwater monitoring, particularly valuable for applications in marine research, environmental monitoring, and security.

Automated feature extraction and segmentation of medical images are essential for accurate diagnostics, enabling the identification of relevant structures with minimal human intervention. This study introduces an Explainable AI (XAI) framework for automated feature extraction in medical image segmentation, aiming to enhance transparency in deep learning models used in medical imaging. The proposed framework uses a Convolutional Neural Network (CNN) with integrated attention mechanisms and layer-wise relevance propagation (LRP) to identify critical features while segmenting regions of interest. Testing on datasets of MRI brain scans and CT liver scans, the model achieved an accuracy of 94%, a Dice similarity coefficient (DSC) of 0.88, and an Intersection over Union (IoU) score of 0.83. These results outperform conventional CNN-based segmentation techniques by 10% on average, highlighting the framework's precision in identifying and segmenting intricate structures, including lesions and abnormalities. Additionally, the XAI components provide visual explanations of the segmentation process, enabling clinicians to understand which features influenced the model's decisions. This enhanced transparency is crucial for building trust in AI-driven medical solutions, ultimately facilitating their integration into clinical workflows.

Medical imaging has become a critical tool in diagnostics, but low-resolution images often limit the precision of diagnosis and treatment. This study presents a deep learning-based image super-resolution framework designed to enhance the quality and clarity of medical images, specifically tailored for radiology, dermatology, and histopathology. The proposed framework uses a Convolutional Neural Network (CNN) architecture with a Residual Dense Network (RDN) backbone, improving visual details and retaining clinically relevant features. Training on a diverse dataset of MRI, CT, and X-ray images, the model achieved a 35% improvement in Peak Signal-to-Noise Ratio (PSNR) and a 42% improvement in Structural Similarity Index Measure (SSIM) compared to conventional interpolation techniques. Our method also demonstrated an increase of 48% in diagnostic accuracy when integrated into radiological workflows, enhancing radiologists' ability to identify pathologies with subtle visual indicators. Experimental results show that our super-resolution framework provides a fourfold increase in resolution while minimizing computational cost by 30% using optimized GPU-based processing. This innovative approach to super-resolution has the potential to significantly impact the diagnostic field by enabling clearer and more detailed medical imaging for improved patient outcomes.