Alzheimer Disease Progression Forecasting: Empowering Models Through hybrid of CNN and LSTM with PSO Op-Timization

A common neurodegenerative disease, Alzheimer Disease (AD) affects society. Early intervention and personalized care require accurate condition prediction. A hybrid model using Convolutional Neural Networks (CNN) and Long Short-Term Memory networks (LSTM) and Particle Swarm Optimization (PSO) is developed in this study to optimize performance. This research uses a large MRI dataset with important neuroimaging data. This study uses a dataset to train and validate our models, enabling a data-centric approach to AD progression. Forecasting involves predicting future events or outcomes using available data. AD causes cognitive decline and memory loss, making healthcare more complicated. Timely prognosis is essential for prompt interventions and personalized patient care. Conventional forecasting models like CNN and LSTM are good at predicting disease progression. CNN excels at capturing spatial dependencies in datasets, while LSTM excels at temporal sequences. We proposed a novel hybrid model to take advantage of both architectures. This paper uses Particle Swarm Optimization (PSO), an effective optimization algorithm, to fine-tune hybrid model parameters. The goal is to improve model forecasting accuracy. In this study, the hybrid CNN-LSTM model with and without PSO accurately predicted AD progression. Our analysis includes accuracy, precision, recall, F1-Score, and ROC AUC to assess model efficacy. This study advances predictive analytics in healthcare and offers new ways to improve AD patient outcomes.