Multi-modal Biometric Emotion Recognition Using Classifier Ensembles.

We introduce a system called AMBER (Advanced Multimodal Biometric Emotion Recognition), which combines Electroencephalography (EEG) with Electro Dermal Activity (EDA) and pulse sensors to provide low cost, portable real-time emotion recognition. A single-subject pilot experiment was carried out to evaluate the ability of the system to distinguish between positive and negative states of mind provoked by audio stimuli. Eight single classifiers and six ensemble classifiers were compared using Weka. All ensemble classifiers outperformed the single classifiers, with Bagging, Rotation Forest and Random Subspace showing the highest overall accuracy.