Using Machine Learning to Build a Model or Early Prediction of Alzheimer's Disease

Abstract

Alzheimer's disease is a neurodegenerative disease and a common factor of dementia in the elderly, and based on the fact that the appearance of its apparent clinical symptoms is late and there is no treatment for it, work was done in this study on the early prediction of Alzheimer's disease by detecting the stage of mild cognitive impairment experienced by most Alzheimer's patients. This study was used the ADNI database. After preprocessing, preparation, balancing and division into 80% as a training set and 20% as a test set, Two models were Hyperparameters tuned, trained and tested.

Two models of Machine Learning were worked on in this study: the first model KNN (K-Nearest Neighbors) and the second model AB (AdaBoost) to classify three cases: Normal Cognition (NL), Alzheimer's Disease (AD) and mild Cognitive improvement (MCI).

The AB model showed higher efficiency than the KNN model, where it outperformed, the Accuracy value of the AB model was a ratio 92.86%, and the Sensitivity value was equal to 92.86%, and these are good results to help support the doctor's decision.