Chronic Kidney Disease is a progressive medical ailment of growing global health importance because, in most cases, this ailment shows no symptoms during its early stages. Improving patients’ outcomes and early detection are significant aspects of managing diseases. In this paper, deep learning models to classify the images of kidney diseases are presented based on a dataset of 12, 446 images collected from various renal diseases. Therefore, CNN, VGG16, MobileNet V2, DenseNet 121, and ResNet 50 were the fine-tuned and evaluated models. The training setting was the Adam optimizer, categorical cross entropy loss, and 10 epochs. Hence, the model's performance was measured using the accuracy, precision, recall, and F1-score evaluation parameters. Following that, the current evaluation illustrates that most of the examined models positively predict outstanding accuracies, with ResNet 50 having a maximal validation and test accuracy rate reaching 99.40%. At the same time, MobileNet V2 and DenseNet 121 also boast of their high efficacy. The researchers' works highlighted that deep learning algorithms are very helpful for diagnosing kidney diseases based on medical images, underlining that their application can significantly change early diagnosis and patient treatment.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style

Deep learning models provide tremendous support for medical imaging by understanding lung conditions and indicating multiple lung diseases. Due to the global burden of respiratory diseases, their prevention and control is of great importance. Therefore, this study focuses on the effectiveness of different deep learning architectures in diagnosing lung diseases from chest X-ray images. Five deep convolutional neural networks are involved: VGG16, DenseNet-121, ResNet-50, MobileNet, and Vision Transformers. They are pre-trained using the ImageNet dataset. Both transfer learning and development of custom models based on the above architectures will be applied. The study deals with the determination of the most effective single model for the identification of lung diseases. The gradient-weighted class activation map is used to highlight the key regions that influence model decisions. In addition, soft voting ensemble learning methods are used to improve the performance of lung disease detection. Commonly used metrics are applied to evaluate all models. The results for COVID-19, pneumonia and normal case identification exceeded 95% accuracy, 95% precision, 96% recall and 95% Fβ for individual models. The ViT model outperformed DenseNet-121, achieving 96.66% accuracy. The results for bacterial pneumonia, viral pneumonia, tuberculosis, COVID-19 and healthy case identification exceeded 85% accuracy, 86% precision, 85% recall and 94% Fβ for single models. Ensemble learning further improved performance. These results demonstrate the high potential of deep learning and ensemble approaches to support accurate and efficient diagnosis of lung diseases using chest X-rays. The deep learning models provide promising decision support tools for this type of healthcare diagnosis.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style

The failure of solder joints of Ball Grid Array (BGA)Package under drop impact is influenced by multiple parameters, highlighting the need for optimization during the early design stages of electronic systems. In this paper, ensemble methods were developed to predict failure in solder joints by estimating the dynamic responses of printed circuit board assembly (PCBA) during drop impact conditions. Finite element (FE) simulations were carried out by varying PCB thickness, PCB modulus, solder ball diameter, and solder ball material to obtain the dynamic responses of the PCBA during impact loading, which served as the dataset for the predictive model. Also drop test experiments were conducted according to the JESD22-B111A standard to validate the FEM results. XGBoost regression achieved the best performance with an R² of 0.96 and the lowest error, with feature importance analysis identifying solder ball material (score: 0.447) as the most influential factor and PCB modulus (score: 0.065) as the least.The predictive model developed in this work offers a robust tool for evaluating mechanical performance and optimizing design parameters in PCBA structures under dynamic mechanical stresses.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style

This research aims to use a vibration monitoring system along with machine learning techniques to predict the downtime and Remaining Useful Life (RUL) of three-phase induction motors in the manufacturing sector. The study obtains measurement data from accelerometer sensors that collect various parameters related to motor performance. The research includes a data preprocessing stage to handle missing data, select predictor attributes, and remove duplicates. Supervised learning algorithms are applied, including Decision Tree (DT), Naive Bayes (NB), Random Forest (RF), and Artificial Neural Network (ANN). The results show that DT and NB models have the best performance in downtime classification, achieving 100% accuracy, recall, precision and F1 values. In terms of predicting Remaining Useful Life (RUL), the RF model outperforms the base model and ANN, showing better results in Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and correlation coefficient.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style

Maternal mortality remains a critical challenge in Sub-Saharan Africa, with Nigeria ranking among the countries with the highest rates. The loss of women in their reproductive years destabilizes families causing emotional trauma, places additional strain on healthcare systems, and has profound economic and national developmental consequences. As a result, one of the United Nations Sustainable Development goals (SDGs) is targetted at reducing maternal mortality and morbidity at all cost. This study explores the application of Artificial Intelligence (AI) in healthcare through the development of a predictive ensemble model to classify maternal health risks as identifying high risk pregnancies can inform timely clinical decision making that mitigates maternal mortality. Maternal health dataset was sourced from three (3) health centers in Delta State, Nigeria.. Nine supervised machine learning classifiers were utilized, including Linear Support Vector Machine, Gaussian Naïve Bayes, Multilayer Perceptron, Decision Tree, Random Forest, Gradient Boosting Decision Tree, Extreme Gradient Boosting, Light Gradient Boosting Machine, and Categorical Boosting. To enhance predictive performance, the classifiers were combined in an ensemble model. Results showed that the Gradient Boosting Decision Tree achieved the highest accuracy at 90% before upsampling and Random Forest achieved an accuracy of 97% at upsampling. The lowest-performing classifier was Linear Support Vector Machine before and after upsampling. The ensemble model surpassed all individual classifiers, achieving 98% accuracy and precision and over 1% increase in accuracy after upsampling. This study highlights the potential of AI-driven predictive models to optimize healthcare resources and improve maternal health outcomes in Delta State, Nigeria.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style

This study aims to evaluate the user experience (UX) of Enterprise Resource Planning (ERP) system modules in Polish enterprises' information and communication technology-based production environments. The research plan includes quantitative research as part of a doctoral thesis, which will be complemented by qualitative methods such as task-based usability tests, heuristic analysis and in-depth interviews with users. A descriptive research design was employed using an online survey incorporating the User Experience Questionnaire (UEQ) to gather quantitative data. The survey was distributed to a diverse group of respondents, including students, alumni, and production practitioners, to capture perceptions of three screens of the ERP system (General Operations Registration, Personalised Operations Registration, and Employee Panel Operations Registration), which enable the registration of operations. Raw responses from a seven-point Likert scale were transformed into a continuous scale and statistical analyses were conducted to compute descriptive metrics and confidence intervals. The findings indicate that the pragmatic dimensions — namely, perspicuity, efficiency, and dependability — received favourable evaluations, demonstrating robust usability and clear functionality. In contrast, the hedonic dimensions, particularly stimulation and novelty, were rated as neutral to negative. This suggests that, although the ERP modules effectively support routine tasks, they lack innovative appeal and engaging design. Benchmark comparisons revealed that the interfaces generally fell within the lower quartile, highlighting the need for targeted UI/UX refinements to enhance visual attractiveness and user motivation. In conclusion, the study highlights the importance of balancing functional performance with improved aesthetic and hedonic attributes to optimise the effectiveness of ERP systems in production engineering settings.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style

Industrial mobile robots face critical positioning challenges that impact manufacturing efficiency, warehouse automation productivity, and biomedical service delivery. This paper presents a reproducible framework for quantifying odometric drift in differential-drive robots, validated by consumer-grade, low-cost VR tracking. Applications include industrial automation calibration, warehouse logistics management, and precision biomedical device positioning. Through more than 750 automated experimental trials spanning a comprehensive matrix of motor configurations and path geometries, the results show that both path complexity and turn size significantly influence drift patterns. Specifically, routes with higher geometric complexity (12-15 segments) exhibited 22% greater position error than simpler paths. The analysis used advanced metrics such as the Normalized Drift Contribution Index. The results confirm robust, high-resolution drift analysis and provide a low-cost validation tool for robot calibration in manufacturing and medical instrumentation. The work provides actionable insights for optimizing robot programming, calibration, and curriculum design, and establishes a scalable protocol for benchmarking autonomous navigation systems in real-world scenarios. In addition, the methodology enables data-driven decision making for robot fleet management, reducing operational downtime compared to manual calibration methods, while providing quantitative performance benchmarks essential for industrial quality control standards.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style

Wireless Body Area Networks (WBANs) are one of the emerging technologies in the healthcare landscape. It enables the non-invasive collection of physiological data to continuously measure health indicators using a network of miniaturized sensors placed on or under the human body. This paper explores a comprehensive study of WBANs, covering all the basic concepts, including their background information and motivation for development, as well as requirements and issues related to their application scenarios and future directions. The paper elaborates on the exclusive characteristics of WBANs compared to Wireless Sensor Networks. It describes health monitoring requirements and energy efficiency challenges with security and biocompatibility as guidelines for comparison. In addition, the paper also highlights various WBAN communication technologies and their relevance in diverse medical and non-medical domains. This paper identifies the critical comprehensive analysis of interference dynamics and mitigation strategies that remain absent in the literature, along with an exhaustive review of the literature. The research shows that WBANs could have a significant impact on healthcare and other industries, while discussing the technical and ethical hurdles to their wider application.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style

Introduction: Accurate and interpretable arrhythmia detection is essential for timely diagnosis and intervention, especially in medical decision support systems (MDSS). Fuzzy logic, known for its ability to handle uncertainty and improve interpretability, has emerged as a promising approach. Aims: This systematic literature review (SLR) examines the role of fuzzy logic in advancing arrhythmia detection, focusing on accuracy, interpretability, and integration with computational intelligence. Methods: Following PRISMA guidelines, 18 studies published between 2019 and 2024 were analyzed to address four key questions: (Q1) the accuracy and reliability of fuzzy logic systems, (Q2) the effectiveness of hybrid systems combining fuzzy logic with computational intelligence, (Q3) the challenges in developing multi-class fuzzy logic systems, and (Q4) the impact of fuzzy logic on interpretability in MDSS. Techniques such as Adaptive Neural Fuzzy Inference Systems (ANFIS) and hybrid models with neural networks and bio-inspired algorithms were evaluated. Results: ANFIS demonstrated near-perfect accuracy, while hybrid systems improved scalability and overcame the challenges of multi-class classification. Limitations included reliance on benchmark datasets, limited real-world validation, and insufficient focus on Explainable Artificial Intelligence (XAI). Conclusions: Fuzzy logic shows great potential for developing interpretable and robust MDSS for arrhythmia detection. Future research should prioritize advancing XAI, incorporating diverse data sets, and addressing real-world challenges to improve clinical applicability.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style

This study aims to improve the interpretability of brain tumour detection by using explainable AI techniques, namely Grad-CAM and SHAP, alongside an Xception-based convolutional neural network (CNN). The model classifies brain MRI images into four categories — glioma, meningioma, pituitary tumour and non-tumour — ensuring transparency and reliability for potential clinical applications. An Xception-based CNN was trained using a labelled dataset of brain MRI images. Grad-CAM then provided region-based visual explanations by highlighting the areas of the MRI scans that were most important for tumour classification. SHAP quantified feature importance, offering a detailed understanding of model decisions. These complementary methods enhance model transparency and address potential biases. The model achieved accuracies of 99.95%, 99.08%, and 98.78% on the training, validation, and test sets, respectively. Grad-CAM effectively identified regions that were significant for different tumour types, while SHAP analysis provided insights into the importance of individual features. Together, these approaches confirmed the reliability and interpretability of the model, overcoming key challenges in AI-driven medical diagnostics. Integrating Grad-CAM and SHAP with a high-performing CNN model enhances the interpretability and trustworthiness of brain tumour detection systems. The findings underscore the potential of explainable AI to improve diagnostic accuracy and encourage the adoption of AI technologies in clinical practice.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style

This study presents a comprehensive noise source analysis of a nitrogen generation system installed in an industrial production facility. The primary objective of the investigation was to determine the location of the dominant noise sources and to identify their respective sound pressure levels and frequency characteristics under operating conditions. Detailed measurements were made using a 16-microphone array combined with CAE Noise Inspector software for accurate sound field visualization and analysis. Experimental tests were conducted at three different locations within the system: the nitrogen generation unit (location 1), the nitrogen storage tank (location 2), and the exhaust pipe (location 3), with the latter further subdivided into three specific measurement points (3a, 3b, 3c) to account for variations along the length of the pipe. Each acoustic measurement session lasted three seconds, with data recorded at a high recording frequency of 204,800 Hz to ensure precise resolution across the frequency spectrum. The operating cycle of the nitrogen generator was divided into two main phases: phase 1, characterized by the transient sounds associated with valve actuation, and phase 2, dominated by the continuous sounds generated during nitrogen transfer to the storage tanks and exhaust. Recordings at site 1 captured both operational phases, while measurements at sites 2 and 3 focused exclusively on phase 2 in order to isolate relevant noise sources. The results provide a detailed and quantitative characterization of the acoustic emissions associated with the nitrogen generation process, providing valuable insights that can be used to develop targeted noise reduction strategies and contribute to future optimization of the system's mechanical design and operational efficiency.

• APA 7th style

• Chicago style

• IEEE style

• Vancouver style