Comparison of Classification Systems for Detection of Drug Effects on the Brain Using Machine Learning-Based EEG Signals

Drug abuse is very dangerous for the body, especially on brain performance that causes brain damage. Drug abuse is a serious global health problem with severe consequences, requiring comprehensive prevention and rehabilitation strategies. Early detection is essential and electroencephalogram (EEG) analysis has enormous value. In this study, drug detection using EEG system with Naïve Bayes and Random Forest classifiers is proposed. In the experiment 40 subjects in one of the prisons in West Java were involved. Based on the experiments conducted, Random Forest provides higher accuracy, which is 93.75%, as evidenced by K-Fold Cross Validation, while Naïve Bayes is only 92.75%. In addition, Random Forest also performed better in other evaluation metrics, such as precision, recall, and Fl-Score, which were higher than Naïve Bayes. These results show the proposed technology using Random Forest classifier has the potential to be used for drug detection.