Evaluation of machine learning models for the prediction of Alzheimer's: In search of the best performance

Abstract

Alzheimer's is a progressive and degenerative disease affecting millions worldwide, incapacitating them physically and cognitively. This study aims to perform a comparative analysis of Machine Learning models to determine the model with the best performance in predicting Alzheimer's disease. The models used were Random Forest (RF), Adaptive Boosting (AdaBoost), Support Vector Machine (SVM), K-nearest Neighbors (KNN), and Logistic Regression (LR). Two datasets called OASIS were used to train the models, the first one had a total of 436 records and 12 variables, while the second one stored 373 records and 15 variables. The article's content is divided into six main sections: introduction, literature review, methodological approach, results, discussions, and conclusions. After processing and pooling the datasets, RF, SVM, and LR proved the best predictors, achieving 96% accuracy, precision, sensitivity, and F1 score. This study highlights the efficacy of RF, SVM, and LR in predicting Alzheimer's disease, offering a significant advance toward understanding and management of this disease, which supports the relevance of implementing these models in future research and clinical applications.