结合相似度的新型加权网络模型

New Weighted Network Model Combining Similarity

许多实际复杂网络都可以采用加权网络模型描述.现有加权网络多以节点强度作为择优连接的概率,而未考虑节点之间内在属性的相似也会增加两个节点连接的概率.基于典型的BBV加权网络模型,提出一个结合相似度的新型加权网络模型,改进了已有模型的连边增长方式和择优连接机制,提出了一种权重自适应演化机制.通过提出节点之间相似度的概念,网络演化中同时兼顾节点强度与相似度进行择优连接.网络增长时,既考虑了新节点与已有节点之间增加连边,又考虑到两个已有节点之间增加连边.理论分析和实验结果表明,该网络模型具有无标度特性和小世界特性,节点度和节点强度均具有幂律分布规律,具有更广泛的应用场景.

Many practical complex networks can be described by weighted network model.The node strength is mostly taken as the probability for preferred connection in existing weighted network,but the similarity of inherent attributes between nodes will increase the possibility of connection between two nodes.Based on the classical BBV weighted network model,a new weighted network model combining similarity is proposed,which improves the edge growth mode and preferred connection mechanism of existing models.Meanwhile,a weight adaptive evolution mechanism is proposed.By proposing the concept of similarity between nodes,both the node strength and the similarity of nodes are taken into account to carry out optimal connection in network evolution.When the network grows,it considers the increase of connections between the new node and the existing nodes,and also considers the increase of the edges between the two existing nodes.The results of theoretical analysis and experiment show that this network model has the property of scale-free and small-world,node degree and node strength both follow power law distribution,which has more extensive application scenarios.