基于共同邻居邻域拓扑稠密性加权的链路预测方法

链路预测旨在利用已知的网络节点和拓扑结构信息,预测网络中未连接的两个节点之间存在连边的可能性。基于网络拓扑相似性的链路预测方法计算复杂度低且预测效果好,但现有的相似性指标对共同邻居的邻域拓扑信息考虑较少。针对此问题,提出一种基于共同邻居邻域拓扑稠密性加权的链路预测方法。首先,基于邻域拓扑相对稠密指数量化节点的邻域拓扑结构;然后,利用共同邻居的节点度和邻域拓扑相对稠密指数刻画共同邻居及其邻域拓扑的相似性贡献;最后,提出基于共同邻居邻域拓扑稠密性加权的节点相似性指标。在多个实际网络数据上的实验结果表明,与现有相似性指标相比,该方法能够取得更高的预测精度。

Link prediction method based on topological density weighting of common neighbor neighborhood

Link prediction aims to use known network nodes and topology information to predict the possibility of edges between two unconnected nodes in the network.The link prediction method based on network topological similarity has low computational complexity and good prediction effect,but the existing similarity indices take less consideration of the neighborhood topological information of common neighbors.To solve this problem,this paper proposed a link prediction method based on the weighted neighborhood topological denseness of the common neighbor.Firstly,the method quantified the neighborhood topology of nodes based on the relative density index of neighborhood topology.Then,it used the node degree of the common neighbor and the relative density index of the neighborhood topology to describe the similarity contributions of the common neighbor and its surrounding topology.Finally,it proposed a node similarity index based on the weighting of the topological denseness of common neighbors.The experimental results on multiple actual networks show that the proposed method can achieve higher prediction accuracy compared with existing similarity indices.