探测和评估复杂网络影响力节点的路径多样性核度中心方法

A method for identifying and ranking influential spreading nodes in complex networks based on neighborhood core diversity centrality

针对当前复杂网络影响力节点探测和评估方法不能精确定位影响力节点、计算复杂等不足,在传统网络K核分解方法的基础上引入了路径多样性概念,从信息传播角度进行了研究,提出了一种基于路径多样性信息熵进行影响力节点探测与评估的新的核度中心方法,即路径多样性核度中心(C_(ncd))方法。实验结果显示,相对于其他影响力节点探测与评估方法,如度中心法(C_D)、介数中心法(C_R)、接近中心法(C_C)、K核中心法(K_C)及核度中心法(C_(ncd)),C_(ncd)方法能够更精确地对影响力节点进行定位,并且能更细粒度地对节点影响力进行有效排序。

The study aimed to find an improved technique for identifying and ranking influencial nodes in complex net- works. In consideration of current techniques＇ problems of lower node locating accuracy, higher computing com- plexity, etc. , a method to desect and extimate influential spreading nodes based on the information entropy of path diversity, called the neighborhood core diversity centrality （Cncd） method, was put forward by introducing the con- ception of path diversity into the traditional K-core decomposition method to conduct the study from the perspective of information propagation. The experimental results show that, compared with other methods like degree centrality （ Co ） , betweennes centrality （CB） , closeness centrality （ Cc） , K-core centrality （ CKC ） and neighborhood core cen- trality （Cnc）, the proposed Cncd method can identify influential nodes more accurately and rank influential nodes more finely.