Detection of Alzheimer’s Disease using Machine Learning Classification

Alzheimer’s disease is a progressive and and irreversible brain disorder that basically affects memory, thinking, and behavior. Overtime, people with this disease experience confusion, difficulty with language and changes in personality. Detecting Alzheimer’s is a complex and time-consuming task that requires brain imaging reports and human expertise. The main aim of the proposed work is to develop a smart and reliable way to diagnose disease at an early stage. Early diagnosis enables timely intervention and treatments which results in potentially slowing down progress of the disease. Identifying individuals with MCI (Mild Cognitive Impairment) who are about to get migrated into Alzheimer’s disease is important in both clinical trials and practices. Combining MRI (Magnetic Resonance Imaging) data and neuro-psychological test results as inputs to a multi-dimensional space for the classification of disease is a promising approach. This approach can lead to several benefits and ultimately resulting in better decisions and significant savings. In the proposed work, Multiple datasets Alzheimer’s disease Neuroimaging Initiative (ADNI) and the Open Access Series of Imaging Studies (OASIS) are used. A variety of supervised machine learning algorithms have been applied for automatic Alzheimer’s disease detection. The performance metrics with respect to the accuracy values have been analyzed and from the results, it is observed that the minimum accuracy of 67.75% is achieved in AdaBoost and a maximum accuracy of 80.22% has been achieved using Gradient Boosting Classifier.