<![CDATA[Stress is a common response to life's pressures that, if left untreated, can negatively impact physical and mental health. Accurately detecting and classifying stress levels is a significant challenge. Electroencephalography (EEG), as a non-invasive method, is capable of recording brain activity and representing a person's emotional state, including stress levels. However, the complexity of EEG data requires effective classification techniques. This study aims to develop a stress level classification system based on EEG signals using the Decision Tree C4.5 method. The EEG dataset was taken from Binjai Prison, with inputs in the form of brain waves (Delta, Theta, Alpha, Beta1, Beta2) and values from various electrodes. The output is a stress classification into three categories: stressed, relaxed, and neutral. The results show that the C4.5 method is able to classify stress levels with 98.68% accuracy, an average precision of 99.19%, and an average recall of 96.29%. The beta2 feature is the most dominant attribute, followed by theta and beta1. Thus, the C4.5 method shows good performance and provides clear interpretation in classifying stress levels based on EEG signals.]]>
Copyrights © 2025
