Research On Classification Of EEG Emotional Brain Network Based On Undirected Weighted K-Order Propagation Number Algorithm

EEG has a strong correlation with human emotion,which can effectively reflect psychological activities and cognitive behavior,and provide a means for emotion classification research.Emotion is the result of the joint participation and interaction of different brain regions.In order to explore the mechanism of emotion production and reflect the relevance of one brain region to another brain region and the information relevance of neural network,this paper constructs the cross-sample entropy emotional brain network based on the mathematical-statistical relationship between the channels from the perspective of the complex brain network.The measurement of node importance is an important means to study the complex network.Analyzing the node importance of complex network can effectively identify the key nodes in the network.In order to make up for the shortcomings of the traditional evaluation method of node importance,this paper adopts a new evaluation method of network node importance,undirected weighted K-order propagation number algorithm,to analyze the difference of node importance in different emotional brain networks.In order to verify whether the undirected weighted K-order propagation number algorithm can effectively evaluate the node importance of the network.In this paper,three undirected networks,i.e.California highway network,ARPA network and protein network(local),are simulated and analyzed,and the results are compared with those of traditional node importance evaluation methods.The simulation results show that the node importance evaluation model based on undirected weighted K-propagation number is more accurate and effective.It can identify the key nodes in the complex network and reduce the number of nodes with the same degree of importance.In addition,it achieves good results in judging the nodes ranking lower in importance,which provides a new method for the study of node importance in the complex network.Then,this paper applies the above node importance theory to the classification of emotions.Firstly,the node importance of different emotional brain networks is calculated based on the undirected weighted K-order propagation number algorithm.Then,support vector machine and principal component analysis are used to extract the node importance details of different emotional brain networks as the input of machine learning to realize the research of two-category and fourcategory of emotional brain networks.The simulation results show that the node importance of different emotional brain networks has great differences.The number of edge connections of negative emotional brain networks decrease faster than that of positive emotional brain networks with the increase of threshold,and the functional connection of negative emotion in the right brain is more intensive than that of positive emotion.For the two-category research of emotional brain networks,83.6% of the two-category accuracy rate is finally achieved.The results show that it is effective to use the weighted K-order propagation number algorithm to extract the important characteristics of brain network nodes for emotion classification.