基于动态网络的链接分析与预测研究

Research On Link Analysis and Prediction Based On Dynamic Network

复杂网络通过新节点和新链接的增加而随着时间快速演化,使用网络的静态特征难以产生精确的链接预测。通过分析动态网络特征,结合张量分解,提出了一种新的基于动态网络的链接预测方法,并应用于真实的复杂动态网络数据集中对未来链接进行分析预测,实验演示了该方法的优点和效果,取得了较好的预测结果。

Complex networks evolve rapidly over time by adding new nodes and new links, and it is difficult to get an accurate prediction by using the static characteristics of networks. By analyzing the dynamic characteristics of the networks, according to the tensor factorization theory, a new method of link prediction based on dynamic network is proposed, which is applied to the real complex dynamic network data for link analysis and prediction, The experiment demonstrates the advantages and effectiveness of this method, which achieves good prediction performance.