Identification of stress source from eeg signals with trigonometric transformation based features

Stress is a condition that people have been exposed to for many years, triggered in many different situations such as daily life, office environment, and family life, and as a result, the human body shows mental, physical, and psychological reactions. In cases where stress cannot be detected early and the treatment process cannot be started, the risk of many different diseases such as heart attack, stroke, and depression is quite high. For this reason, many researchers are conducting studies on early detection of stress using questionnaires, hormonal tests, and physiological signals. As seen in most of the studies, stress approaches are compared with relaxation state. In this thesis, the classification of different stress sources, which is considered to be a more difficult problem than the comparison of the stress approach according to the relaxation state, was studied. The features of 40 participants' EEG data recorded during 3 different stress sources were generated using common trigonometric transformation functions and the proposed trigonometric transformation functions. The 3 different stressors were divided into binary subclassification problems and classified by 4 different classification algorithms and compared with each other.