Deep Learning Based Hybrid Framework for Early-Stage Prediction of Alzheimer Disease: A Predictive Analysis

Alzheimer Disease (AD) is a progressive neuro-degenerative disorder that impacts millions worldwide, presenting a considerable healthcare issue. Usually, symptoms start out mildly and get worse with time. Early identification is essential for effective treatments and decelerating disease development. Predictable diagnostic methods frequently show constraints in scalability and accuracy, requiring the development of advanced computational systems. This study introduces a Deep learning-based hybrid framework for the identification of reliable symptoms for the early prediction of AD. Deep Learning has become an essential tool in healthcare due to its ability to identify complex patterns in large datasets. This research combines two deep learning architectures: ResNet50 for feature extraction from medical images, such as MRI, and MobileNet for collecting temporal patterns, applied on the Alzheimer's OASIS dataset to detect early-stage Alzheimer's disease. The framework's performance is assessed using important metrics: sensitivity, specificity, and overall accuracy during the training, validation, and testing phases. This study emphasizes the significance of multidisciplinary deep learning methodologies in transforming early diagnosis and the possibility of hybrid deep learning strategies in progressing medical science to address intricate healthcare challenges.