Research On Multifunctional Sentiment Analysis Application System Based On EEG

With the development of biomedicine and computers,brain science has become a hot research direction.EEG(Electroencephalogram)signals are an important feature of brain activity.The rapid development of brain-computer interfaces has caused a large number of researchers to try to uncover the mystery of the brain through EEG signals.Emotion and attention have an important impact on human production and life,and are currently the research hotspots in the field of brain science.This paper designs and implements an application system for detecting human emotions and attention based on EEG signals.The functions include driving state monitoring,class listening quality monitoring,EEG database establishment.The emotion correction based on the multi-core function SVM(Support Vector Machine)voting algorithm.After that,a tensor machine emotion classification algorithm based on synchronous brain network is proposed,in order to further improve the accuracy of the system for emotion classification.First of all,a multifunctional EEG signal analysis embedded system was developed.According to the single-lead EEG signal collected from the prefrontal lobe,it analyzes human emotion,attention and relaxation,and the data obtained from the analysis is applied to three aspects: driving status monitoring,class listening quality monitoring and EEG database establishment.The driving status monitoring function is to monitor the driver’s attention status in real time,and give voice warnings when the driver is not paying attention;the class quality monitoring function is used to monitor the status of students during class,and feedback to the teacher so that they can adjust the teaching method in time.Improve classroom efficiency;the EEG database establishment function is used in the field of scientific research,through the system’s own video stimulation to induce users to produce corresponding emotions,and at the same time collect EEG signals,establish a corresponding database for EEG emotion research.The hardware part of the system includes a wireless single-lead EEG signal acquisition headgear and a portable mobile analysis terminal.The EEG signal collection headgear transmits the collected EEG signals to the mobile terminal for analysis and processing via Bluetooth.The software part of the system includes EEG analysis module,three functional modules and the top-level module.The EEG analysis module is used for signal denoising,feature extraction and model training,and the result after analysis is called by the function module.The top-level module is used for human-computer interaction,and the user selects and calls function modules through the top-level module.In order to further improve the accuracy of system recognition,this paper proposes an emotion classification algorithm Support Tensor Machine based on Synchronous Brain Network(SBN-STM),which uses Phase Locking Value(PLV)to construct a synchronous brain network and analyzes multiple leads The synchronization and correlation between EEG signals are generated,and a second-order tensor sequence is generated as a training set,and the Support Tensor Machine(STM)model is used to realize the two classification of positive and negative emotions.Then based on the DEAP(Database for Emotion Analysis using Physiological Signals)EEG emotion database,the selection method of the synchronous brain network tensor sequence is analyzed in detail,and the size and position of the optimal tensor sequence window are solved.The remaining problems have improved the speed of model training.Simulation experiments show that the emotion accuracy of the synchronous brain network classification method based on support tensor machine is better than that of support vector machine,C4.5 decision tree,artificial neural network,K nearest neighbor and other emotion classification models characterized by vectors.