Advanced Machine Learning Techniques for Alzheimer’s Disease Detection: A PCA and Improved XGBoost Approach

Abstract

Alzheimer’s Disease (AD) is one of the main fields in clinical medicine that contributes to the existing difficulties in research. The presented work is dedicated to the Machine Learning approach to the identification and detection of the Alzheimer’s Disease, including the Image Enhancement techniques. The work of they also use Principal Component Analysis (PCA) together with the contemporary approach in improving the image quality of the brain images obtained from public databases. The focus of this research is the enhanced XGBoost classification model as applied with the help of two other classification methods to assure its effectiveness. A lot of tests were performed on the Alzheimer’s Disease dataset with an analytical feature extraction procedure to enhance the model results. These proposed methodologies are tested against conventional algorithms with an emphasis on accuracy, precision, recall and F1-score. The first estimates suggest an increase in the level of AD detection accuracy and its superiority over conventional approaches. Apart from showing a correlation between PCA and new pre-processing methodologies, this study also underscores the enhanced diagnostic aptitude of the improved XGBoost classifier.