基于节点相似性有偏游走的多层时序网络节点重要性评估

时序网络中关键节点的挖掘引起了社会广泛的关注。针对时序网络建模中存在忽略时间信息、未考虑时间切片间的交互关系进而影响关键节点识别的准确性和科学性的问题，构建了多层时序网络模型，并提出了一种基于节点相似性有偏游走的关键节点识别算法：多层时序有偏PageRank（MTB-PR）。本文中网络模型的构建引入多层网络分析方法，完整涵盖了时序网络的结构演变。同时，基于所构建的网络模型，综合层内相邻节点间相互作用及其层间影响的双重因素来区分节点的不同影响力；通过数据分析，探讨了一些偏差参数变化对节点重要性能指标的影响。最后，将模型和算法应用于真实网络中，通过实验数据验证了该方法的合理性和有效性。

Node Importance Evaluation for Multilayer Temporal Network Based on Node Similarity Biased Walk

The mining of key nodes in temporal networks has attracted widespread attention. Aiming at the problems of ignoring the time dimension and lacking of connection between different time windows in the modeling of temporal networks, which affected the accuracy and scientificity of key node identification, a multilayer temporal network model was constructed and a key node identification algorithm based on biased random walk of node similarity: multilayer temporal biased PageRank (MTB-PR) was proposed. A multilayer network analysis method was introduced in the construction of network model, which completely covered the structure evolution of temporal network. Meanwhile, based on the constructed network model, the dual factors of interaction between adjacent nodes and interlayer factors were combined to distinguish the different influences of the nodes. The influence of some deviation parameter changes on the important performance indicators of the nodes was discussed through data analysis. Finally, the models and the algorithms were applied to real networks, and the rationality and effectiveness of the method were verified through experimental data.