Heterogeneous network interactive visualization method

Abstract

The invention discloses a heterogeneous network interactive visualization method. The method comprises the steps that (1) nodes in a heterogeneous network are clustered according to node attribute values, and a corresponding clustering network visualization map is generated; (2) as for each node in selected clustering results, a neighbor node set of the node is calculated; then according to the node attribute values of the adjacent node sets, the nodes in the clustering results are clustered; a clustering visualization map is generated as a next-level visualization map; (3) the clustering results obtained in the step (2) are selected, the nodes in the clustering results are clustered according to the adjacent node sets of the nodes, and the nodes with the same adjacent node set fall into the same cluster; then the clustering results are used for generating a clustering visualization map which is used as a next-level visualization map of the clustering network visualization map generated in the step (2). According to the heterogeneous network interactive visualization method, topological information and attribute information are combined, and a user can check the lower level with the finer granularity.

Description

technical field [0001] The present invention relates to the fields of clustering analysis, heterogeneous network, data mining, topology analysis, large-scale data network visualization analysis, etc., and proposes a method for processing large-scale heterogeneous network data based on the combination of heterogeneous network node attributes and network topology. method. This method is suitable for typical information network data such as social network, computer network, sensor network and knowledge network. It is a visual display method that can perform interactive analysis. Background technique [0002] With the advent of the era of big data, complex and changeable large-scale data networks have been produced in a short period of time. Many people have begun to analyze and conduct in-depth research on these large-volume and miscellaneous network data. Among these data networks, some are network types with different attributes and characteristics on network nodes, and som...

AI Technical Summary

Problems solved by technology

However, the granularity of the clustering results of these two methods may be too rough to lose a lot of network details (for example, the method based on node attributes compresses the entire network into several node clusters, corresponding to several values ​​​​of node attributes), or Too fine-grained so it is difficult to layout and display (such as clustering based on network structure)

Although there are some methods that can use the number of clusters as an input (such as the segmentation method in the figure), it is difficult for users to control and understand the clustering results, and it cannot support user-defined top-down visual browsing analysis process

Images