Research On Community Detection Base On Hybrid Algorithm Of Beetle Antennae Search And Particle Swarm Optimization

Many complex systems can be represented by complex networks in real life.Community structure is an important characteristic of complex networks.The connection between nodes is close in the same community,and the connection between nodes is sparse in different communities.Detecting community structure can mine the potential structure of the network,reveal the internal law of the network or predict the dynamic evolution trend of the network.It has important theoretical significance and practical value for web community mining,epidemic disease prevention and control,distribution network resource optimization and so on.However,the complex network has a large scale and complex structure,and the boundary between different communities is not clear,which brings challenges to community detection.In view of this,combined with the advantages of beetle antennae search algorithm and particle swarm optimization algorithm,this paper studies community detection from two aspects of overlapping community and non-overlapping community,and applies the community detection algorithm to the personalized film recommendation system.In view of the randomness of existing community detection algorithms based on particle swarm optimization in non overlapping community detection and the lack of accuracy of modular function,starting with population initialization and particle update strategy,a non overlapping community detection algorithm based on hybrid optimization of beetle antennae search algorithm and particle swarm optimization(DBAS-PSO)is proposed.Firstly,the chaotic strategy is used to initialize the algorithm,which improves the quality of the initialization population and reduces the randomness of initialization.Then,the update formula of beetle antennae search and particle swarm optimization algorithm is improved to accelerate the convergence of the algorithm.Aiming at the problem that particle swarm optimization is easy to fall into local optimization,the differential mutation strategy is introduced to make it jump out of local optimization,and the node error correction strategy is added to improve the accuracy of community division.Finally,the algorithm is tested on artificially generated network data and real network data respectively.The simulation results verify the effectiveness of the method in non overlapping community detection.To solve the problem that DBAS-PSO cannot distinguish overlapping communities in the network,,an overlapping community detection algorithm based on hybrid optimization of beetle antennae search and particle swarm optimization(DBPSO-OCD)is proposed.An overlapping node identification method based on neighbor community similarity is proposed to classify the nodes in the network and screen the overlapping nodes in a specific node group,so as to reduce the number of overlapping node identification and improve the efficiency of the algorithm.A community label assignment method based on node similarity and community belonging is designed to assign labels to the selected nodes and divide the overlapping community structure in the network.Finally,experiments are carried out on real network data,and the results show that this method can accurately find the overlapping nodes in the network.The DBAS-PSO community detection algorithm is applied to the recommendation system to realize the personalized film recommendation prototype system.The system makes personalized recommendation for the recommended users by looking for similar user sets.Firstly,the user similarity matrix is calculated through the user’s score on the film,and the network structure between users is obtained.The user network is divided into communities by using DBAS-PSO algorithm to find similar user sets.Finally,personalized video recommendation is carried out based on the preferences of similar user sets for the film.