Diabetes is a chronic metabolic disorder that affects millions of people worldwide, making early detection crucial for effective management. This study assesses the effectiveness of three machine learning (ML) models, Logistic Regression (LR), Naïve Bayes (NB), and Decision Tree (DT), in predicting diabetes based on data from 392 individuals, including their demographic and clinical characteristics. The dataset underwent preprocessing to maintain data integrity, was standardized for model compatibility, and analyzed through feature correlation heatmaps, feature importance assessments, and statistical significance tests. The findings revealed that LR surpassed the other models, with the highest accuracy (78%), precision (73%), and F1-score (65%) for diabetic cases. NB showed moderate performance with 75% accuracy, while DT demonstrated the lowest accuracy (71%) due to overfitting. Receiver Operating Characteristic (ROC) analysis revealed strong discriminative power across all models, although perfect Area Under the Curve (AUC) scores indicate potential overfitting needing further validation. The study emphasizes the significance of key features like Glucose, Body Mass Index (BMI), and Age, which showed notable differences between diabetic and non-diabetic individuals. By enabling early detection and proactive management, these models can contribute to reducing diabetes-related complications, enhancing patient outcomes, and lessening the burden on healthcare systems. Future research should investigate ensemble learning, deep learning, and real-time data integration from Internet of Things (IoT) devices to improve predictive accuracy and scalability.

Brain tumors represent a critical medical condition where early and accurate detection is paramount for effective treatment and improved patient outcomes. Traditional diagnostic methods relying on Magnetic Resonance Imaging (MRI) are often labor-intensive, time-consuming, and susceptible to human error, underscoring the need for more reliable and efficient approaches. In this study, a novel deep learning (DL) framework based on the Densely Connected Convolutional Network (DenseNet) architecture is proposed for the automated classification of brain tumors, aiming to enhance diagnostic precision and streamline medical image analysis. The framework incorporates adaptive filtering for noise reduction, Mask Region-based Convolutional Neural Network (Mask R-CNN) for precise tumor segmentation, and Gray Level Co-occurrence Matrix (GLCM) for robust feature extraction. The DenseNet architecture is employed to classify brain tumors into four categories: gliomas, meningiomas, pituitary tumors, and non-tumor cases. The model is trained and evaluated using the Kaggle MRI dataset, achieving a state-of-the-art classification accuracy of 96.96%. Comparative analyses demonstrate that the proposed framework outperforms traditional methods, including Back Propagation (BP), U-Net, and Recurrent Convolutional Neural Network (RCNN), in terms of sensitivity, specificity, and precision. The experimental results highlight the potential of integrating advanced DL techniques with medical image processing to significantly improve diagnostic accuracy and efficiency. This study not only provides a robust and reliable solution for brain tumor detection but also underscores the transformative impact of DL in medical imaging, offering radiologists a powerful tool for faster and more accurate diagnosis.

Alzheimer’s disease (AD), a progressive neurodegenerative disorder characterized by severe cognitive decline, necessitates early and accurate diagnosis to improve patient outcomes. Recent advancements in deep learning (DL), particularly convolutional neural networks (CNNs), have demonstrated significant potential in medical image analysis (MIA). This study presents a robust CNN-based framework for the classification of AD using magnetic resonance imaging (MRI) data. The proposed methodology incorporates contrast stretching for image preprocessing, followed by principal component analysis (PCA) and recursive feature elimination (RFE) for feature selection, to enhance the discriminative power of the model. The framework is designed to classify MRI into four distinct categories: non-demented, very mildly demented, mildly demented, and moderately demented. Experimental validation on a comprehensive dataset reveals that the proposed approach achieves exceptional performance, with a validation accuracy of 97% and a training accuracy of 100%, alongside reduced loss and improved sensitivity. The integration of PCA and RFE is shown to effectively reduce dimensionality while retaining diagnostically critical features, thereby optimizing the model’s efficiency and interpretability. These findings underscore the potential of DL techniques to revolutionize the early detection and diagnosis of AD, offering a powerful tool for clinical decision-making and advancing the field of neuroimaging analysis. The proposed framework not only addresses the challenges of high-dimensional data but also provides a scalable and generalizable solution for the classification of neurodegenerative disorders.