Scale-free Characteristics and Link Prediction in Complex Railway Network

Abstract

Abstract The link-prediction of complex networks is a new method of network structure mining, which is different from the traditional methods based on machine learning. It reveals those existing relations and predicts possible ones, by using the similarity index of network structure. This paper studied the topology of the railway and link-prediction. The railway dataset was firstly obtained by sampling and pre-processing. Its degree obeys the power-law distribution with the scale-free characteristics. Then, we give some predictions for railway networks. These tests reached good prediction accuracy. The research result has demonstrated that the topological structure of network could be better choice to the link-prediction of a real complex networks.