Alcoholism Detection from Electroencephalogram Using Machine Learning and Deep Learning

Alcohol addiction is a pervasive public health issue with profound societal and individual consequences, necessitating early and accurate detection for effective intervention. This study explores the potential of machine learning (ML) and deep learning (DL) techniques in detecting alcohol addiction through the analysis of electroencephalogram (EEG) signals. By leveraging statistical, frequency-domain, and entropy-based features extracted from EEG data, we implement robust feature selection methods such as recursive feature elimination and T-tests, followed by dimensionality reduction using principal component analysis (PCA) with a cumulative explained variance threshold of 95%. Advanced ML models, including support vector machines (SVM), XGBoost, and K-nearest neighbors (KNN), are compared against Convolutional neural networks (CNN), recurrent neural networks (RNN), and long short-term memory networks (LSTM) are examples of deep learning architectures. Our findings reveal that integrating data-driven ML and DL approaches can achieve high classification performance, offering a non-invasive, reliable, and scalable solution for alcohol addiction detection. This research contributes to advancing computational diagnostics in neurophysiological disorders, paving the way for personalized and timely therapeutic inventions.