Alzheimers Disease Detection through LASSO and RFE based Feature Selection from EEG Data

Alzheimer’s disease is a progressive neurological condition that affects memory, cognition, and basic daily functions, with symptoms typically appearing later in life. This study aims to develop a reliable, non-invasive method for early detection of Alzheimer’s using electroencephalogram (EEG) data. Given the absence of a cure, precise diagnostic tools are crucial. EEG is an affordable and non-invasive alternative, particularly beneficial for disadvantaged populations. Our approach focuses on improving early diagnosis, monitoring disease progression, and providing individualized care. Using Recursive Feature Elimination and Lasso for feature selection, we classified EEG data with machine learning algorithms (Random Forest, Support Vector Machine, and Decision Tree), with Random Forest achieving the highest accuracy of 95.45%.