基于MRLT模型多关系社交网络影响力最大化研究

社交网络影响力最大化问题是基于特定的传播模型，在网络中寻找一组初始传播节点集合，通过其产生最终传播影响范围最大的一种最优化问题。已有的相关研究大多只是针对单关系社交网络，即在社交网络中只存在一种关系。但在现实中，社交网络的用户之间往往存在着多种关系，并且这多种关系共同影响着网络信息传播及其最终影响范围。在线性阈值模型的基础上，结合网络节点间存在的多种关系，提出MRLT传播模型来建模节点间的影响力传播过程，在此基础上提出基于反向可达集的MR-RRset算法，解决了传统影响力最大化问题研究过程中由于使用贪心算法所导致的计算性能较低的问题。最后通过在真实数据集上的实验对比，表明所提方法具有更好的影响力传播范围及较大的计算性能提升。

Influence maximization based on MRLT model in multi-relationships social network

Influence maximization of social network is an optimization problem of finding a set of initial propagation nodes in the network such that the influence range invoked by these nodes is maximized. Most of the existing works have focused on single relation social networks, namely, there is only one relationship in social networks, but in reality, there are a variety of relationships between users of social networks, and these relationships affect the propagation of network information and the influence scope. This paper proposed the MRLT information propagation model, which based on the linear threshold model, combined with the various relationships existing between network nodes, and modeled the influence between nodes. And it proposed the MR-RRset algorithm based on reverse reachable set to solve the problem of low computational performance caused by greedy algorithm in the process of traditional influence maximization research. Finally, the experimental comparison on real data sets demonstrates that the proposed method has better influence propagation range and larger performance improvement.