A Study Of Emotional Ontology With Bayesian Network

Recognizing the emotion context with human bio-signal is highly demanded in nowadays applications, however, the current emotional ontology cannot handle the probabilistic information in emotion recognition process. The primary goal of this research is to draw the Bayesian Network into the study of EEG-based emotion recognition to address the probabilistic context data.The uncertainty of the Bayesian network inference method is introduced to the probability of information based on the EEG (EEG) to identify the emotional state of ontology modeling, on the basis of the original OWL ontology model extended support uncertainty reasoning, which let the emotional data uncertainty modeling and reasoning. The research paper using the Bayesian network model, Affective multi-modal signal database DEAP in EEG and user scale data, verify the accuracy of the Bayesian network constructed to identify the emotional state and Using this Bayesian network modeling and inference rules of the uncertainty extract. Ultimate realization of the Bayesian network, an EEG can determine the emotional state on the basis of the Bayesian network to extend the prototype ontology "Emotiono" function on the uncertainty reasoning.In addition, when using Bayesian network classifiers, in the two dimensions of emotional Arousal and Valence, the classification accuracy rate of the EEG data for the emotional state in two dimensions, respectively,86.8%and85.9%, with higher classification level.