Metody pro analýzy emocí z biologických signálů

Abstract

The diploma thesis is focused on the processing of heart rate variability (HRV) signals and especially galvanic skin response (GSR) in order to classifying subjects according to meditation results. The theoretical part of the thesis contains basic information about GSR and HRV signals and their connections with emotional state. Furthermore, data acquisition and their preprocessing processes, appropriate features that describe the properties of the signals, features extraction algorithms, methods of features classification and evaluation of classification results are mentioned there. In the practical part, a continuous decomposition analysis of GSR signals by using SW Ledalab is performed and features are calculated from the decomposed signals. Based on the features, Fuzzy C-Means clustering is used to classify individual measurements and subjects. The aim of the thesis is the time analysis of individual measurements, the classification of each subject into clusters according to meditation results and definition of each cluster. The analysis of HRV signals is a part of the practical part only marginally.