Early Stage Prediction of Alzheimer's Disease via Machine Learning Models

Abstract

The primary cause of dementia in older individuals is Alzheimer's disease (AD). Metabolic conditions such as Alzheimer's disease and diabetes are prevalent worldwide and are currently under investigation using machine learning techniques due to their increasing incidence rates. Alzheimer's disease is a neurodegenerative condition that affects the brain, and as our population ages, the impact on memory and overall functionality will affect more individuals, their families, and the healthcare system. These consequences will have significant social, financial, and economic implications. Detecting Alzheimer's disease in its early stages is challenging but crucial. Early treatment is more effective and minimizes the extent of damage compared to treatment at later stages. Consequently, this paper proposes an ensemble approach to identify Alzheimer's disease at an early stage. Various methodologies, including ensemble model classifiers, random forests, decision trees, and support vector machines, have been utilized to determine the most effective parameters for predicting Alzheimer's disease. Using Open Access Series of Imaging Studies (OASIS) data, the performance of these machine learning models is assessed using metrics such as precision, recall, accuracy, and F1-score. This classification approach can assist clinicians in diagnosing these disorders. Reducing annual Alzheimer's disease mortality rates is a highly desirable outcome. The proposed method for early disease detection achieves an accuracy of 94% on AD testing data, which is noticeably higher than that of previous studies