| In recent years,online social networks represented by Twitter,Weibo and Facebook are developing rapidly.The increasing availability of online data promotes social networks analysis and mining research.Online social networks(OSNs)provide a new platform for product promotion and advertisement.Influence maximization problem arisen in viral marketing has received a lot of attentions recently.Signed networks are social networks that can reflect positive and negative relationships between users.However,previous studies have rarely considered the friendly or hostile relationship between users,that is,ignored the polar relationship between users,which is inaccurate in practical scenarios.When selecting seed users in signed networks,in addition to considering their positive effects,their negative effects also play a role that cannot be ignored.Therefore,in signed networks,the problem of maximizing the net positive influence is studied.The specific researching contents of this paper are as follows:For signed networks,the problem of maximizing the net positive influence considering user wishes is proposed.The problem can be described as follows:using a signed social network to portray a social network with friendly(positive)and hostile(negative)relationships between people,every person has his own willingness for the information spreading.Select k users from the network,so that make the net positive influence on users the most.Through the detailed analysis of the problem,propose a propagation model at first,then prove that the problem is neither monotonic nor submodular under the model.At last propose a algorithm based on probability driven structure-aware algorithm.Experiments on three datasets show that the seed set found using the proposed algorithm has a better net positive influence.Time is also an important factor in the process of information spreading.For signed networks,the problem of maximizing net positive influence considering time limit is proposed.Select k users from the network in fixed time,so that make the net positive influence on users the most.Through the detailed analysis of the problem,propose a propagation model at first,then prove that the problem is neither monotonic nor submodular under the model.At last propose a algorithm based on simulated annealing algorithm.Experiments on three datasets show that the seed set found using the proposed algorithm has a better net positive influence and high efficiency. |
