Link Prediction In Dynamic Networks Based On Machine Learning

One of the typical features of complex intelligent systems is the non-linear and constant interaction process between components. In order to explore the potential relationships and the evolution pattern of interaction between components, in this paper these systems are modeled as complex networks with dynamically generated links. Then the characteristic time series of dynamic networks are established on the basis of network topological characters and generation times of links A link prediction method for weighted dynamic networks is proposed by combining statistical model and supervised learning method. The experimental results based on real network datasets show that compared with traditional static link prediction methods and unweighted dynamic link prediction methods, the method proposed in this paper can improve the prediction accuracy to a certain extent.