基于动态网络模式变化的网络结构增强嵌入

Network Structure Enhancement Embedding Based on Dynamic Network Pattern Change

动态网络是网络结构和节点属性会随着时间推移而发生变化的复杂网络。由于动态网络能够更好地表达现实世界,从而引起研究者们的兴趣。现实中的动态网络常常会出现网络模式的变化,但是现有的动态网络模型在嵌入时往往将所有的快照划归为同一模式来学习,并没有特别区分其中隐含的网络模式。针对这一情况提出了网络结构增强嵌入(NS-PCN)框架,该框架根据网络模式的变化,能够有效地提取网络中的模式信息。具体表现为,该框架首先判断动态网络时间片所处的模式;然后对相同模式网络之间的网络结构信息增强;最后在4个真实数据集上进行了链路预测实验。实验结果表明,现有的动态网络嵌入模型在本框架下,网络嵌入效果有显著地提高。

Dynamic networks are complex networks as their structures and node features change over time.However,they can better represent the real world,thus attracting the interest of researchers.Although realistic dynamic networks often exhibit changes in their patterns,the existing dynamic network models tend to classify all the snapshots as having the same pattern to learn during their embedding.The NS-PCN framework is proposed for this situation,which can extract the pattern information in the network efficiently according to the change of network patterns.Finally,link prediction experiments are conducted on four real datasets,and the obtained results show a significant improvement of the existing dynamic network embedding model in the present framework.