Application was created for my Master Thesis.
It gives possibility for:
- prepare data structure
- clean data
- preprocess data (remove artifacts of EEG signal)
- feature extraction (using PyEEG)
- feature selection (using Sklearn)
- create classifiers
- prepare plots (statistics and comparison of classifiers)
Used libraries:
Some of important files are hidden like psychological data from participants.
Emotion classification using EEG signals.
Studies of emotions in recent years have gathered a lot of attention in the field of technology. This work is based on the subject of emotion recognition using wearable devices, focusing on the EEG signal. The study was conducted on 43 participants, of which two had to be excluded. During watching videos that evoked emotions, the subjects were equipped with devices for measuring physiological signals, such as EEG, HRV, and EDA. After each video, the participants had to complete questionnaires to determine the intensity of experienced emotions. The videos were displayed and assessed on a specially developed for this purpose application for the Android platform. Signal was preprocessed and features were extracted.
The binary and multiclass classification (nine class) was carried out using various methods. For the analysis of the findings, error matrixes, and statistical charts were prepared, presenting the evaluation of participants' emotions for a given category of video. Their interpretation shows that viewed films did not evoke uniform emotions and often activated several of them simultaneously. This thesis examines various research problems and investigates the different proceedings in the classification process.
Contact if you have questions.
Thanks!