Community-based Analysis Of The Propagation Of Influence In Social Networks

Finding a wide range of influential users from the social network plays an important role in many areas,such as innovation adoption,diffusion and guidance of public opinion,formation and development of group behaviors.Through this transmission of influence in social networks,commercial marketing create greater social influence and commercial value,with promoting new products to the whole market at a lower cost.The discovery of the influential Top-K node set from the social network is of great significance to the research of rapid propagation of information on the network.The discovery of the influential Top-K node set is selecting the set of users of size K from the network and using this set as the trigger set to maximize the spread of information in the network.However,the maximization of influence is NP-Hard problem.Although the classical greedy algorithm can get well approximate result of the optimum,the computational cost of the algorithm is too high.The community structure in the network has a feature that a strong connection between users within the communities and weak connection outside the communities,meanwhile the speed of information dissemination will be affected by the feature.Therefore,according to the characteristics of the community in the network,this article will import the time efficiency of the influence algorithm by using the comprehensive analysis method of network topology information and user behavior under the premise of guaranteeing the high approximate solution.The main work of the paper is as follows.1、Because of the idea of LeaderRank algorithm,CLR(Community-based-LeaderRank)algorithmbased on community structure is proposed to explore the influential Top-K node set in the single network.The idea of CLR algorithm is based on the characteristics of the LeaderRank algorithm to show the characteristics of the community,and then select the influential Top-K node set by calculating the topological properties of the network.2、In the multi-network,the SMCLR(Super-Multi-Community-based-LeaderRank)algorithm combining the community structure is proposed,extending the CLR algorithm of single network to the multi-network and using the network integration analysis method and the collaborative analysis method,research propagation of the influence of multi-network.In this paper,the results of experiment on Twitter datasets show that the performance of combination of community influence is better to find the influential Top-K node set than traditional algorithm.Whether it is CLR algorithm in single network or SMCLR algorithm in multi-network,the influence of selecting Top-K set is almost the same as that of the traditional greedy algorithm,the CLR algorithm and SMCLR algorithm is far superior to greedy algorithm in the time efficiency.