Exploring Influence Maximization On Cross Propagation In LBSN

The Online Social Network(OSN)has become very popular,such as Weibo,WeChat,Facebook,Twitter,etc.People share what is happening around them based on mutual communication.The huge amount of data generated by online social network platform brings unprecedented opportunities to social network analysis.This has attracted many researchers to study the issues,such as the structure of social networks,the law of communication,the analysis of public opinion,etc.The influence maximization is one of the key issues in social network.Most of the current studies focus on online social network while ignoring the offline interpersonal relationship network.Fortunately,the cross propagation considers the characteristics of both the online social network and offline interpersonal relationship network,which is more suitable for the real scenarios.In this paper,we design a cross propagation model based on location-based social network to establish a connection between the online social network and offline interpersonal relationship network.The offline interpersonal relationship network is built by mining the encounter characteristics.Then,an influence maximization algorithm is proposed based on cross propagation model to maximize the influence.The two actual datasets named Brightkite and Gowalla are investigated in this paper.The simulation results indicate that the propagation effect of influence with cross propagation model is better than that in online social network,and the proposed algorithm has higher performances in terms of the precision and sphere of influence.