Abstract: Most of the existing node importance ranking methods only study the topology of the network， ignoring the attribute information contained in the network nodes themselves. Attribute information is very important， but there is widespread incompleteness. Incomplete attribute information is closely related to the importance of nodes. Aiming at these problems， a network node importance analysis method based on superior rough set theory and TOPSIS method is proposed， which combines network structure characteristics and node attribute information and overcomes the limitations of single topological structure analysis. Finally， the proposed method is applied to the user importance evaluation in Weibo social network and compared with other methods. The results show that the ranking results of this method can well synthesize the importance of each node in attribute information and structural characteristics and can fully reflect the importance of each node.

Key words: complex network; node importance; Weibo network; attribute importance; structure importance