PERSONAL THERMAL COMFORT PREDICTION BASED ON EEG SIGNAL

Quantitative measurements of thermal comfort conditions are required for a more valid measurement result than using a questionnaire only. This research aims to conduct a preliminary study using electroencephalography (EEG) signals to predict personal thermal comfort in an indoor environment. The individual's satisfaction or dissatisfaction describes personal thermal comfort to the thermal condition exposure. The applied classification method in this research is the k-Nearest Neighbor classification. The obtained results show that the brain's occipital lobe (represented by the O2 channel) and the frontal lobe (represented by the FC5 channel) were suspected can quantizing personal thermal comfort. The quantization was generated in the delta (0-4 Hz) and theta (4-8 Hz) frequency ribbon in the O2 channel, as well as the beta (13-30 Hz) frequency ribbon in the FC5 channel. With its accuracy of 85%, the k-Nearest Neighbor algorithm was suitable to predict personal thermal comfort.