The Classification Method Based On The Shortest Path Between Nodes Of Network Data

This paper mainly studies the classification of the network data.We propose a new classification method,which is called the classification method based on the shortest path between nodes.For the data produced by the random block structure model,we compare the new method with the Laplace method given by Chung(1997).Simulations results show that the new method has some advantages compared with the Laplace method.Secondly,the problem to choose the number of classification is important,so we give a method based on information entropy.