Topological data sampling method and device, visualization method and system

Abstract

The invention provides a topological data sampling method and device and visualization method and system. The topological data sampling method comprises the following steps: obtaining total out-degreeof nodes in a topological data network structure and maximum node out-degree, and determining whether the topological data network structure is a broadcast topological network or a chain topologicalnetwork according to the ratio of the maximum node out-degree to the total out-degree; when the topological data network structure is the broadcast topological network, determining a first sample nodeaccording to the out-degree of each node in the topological data network structure, and carrying out sampling according to a branch where the first sample node locates; and when the topological datanetwork structure is the chain topological network, determining a second sample node according to the out-degree and level of each node in the topological data network structure, and carrying out sampling according to a branch where the second sample node locates. The sampling method effectively filters the noise data in an original network and keeps overall skeleton of the original network, and ensures a visualization effect on the basis of simulating the topological structure of a complete network.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a topological data sampling method and device, and a topological data visualization method and system. Background technique [0002] In recent years, social software has developed rapidly and has become an indispensable communication tool for users. A commonly used social software can have more than 800 million monthly active users, and the behavior data of these users has formed a very large-scale and complex topological data network. Limited by technology and cost, it is difficult to directly visualize such a large-scale and complex topological data network, so it is difficult to obtain complete information of the network. How to reasonably sample the topological data network and obtain the topological structure that can simulate the complete network Sampling networks to enable accurate inference of properties of parts of complete networks is of great value. [0003] ...

AI Technical Summary

Problems solved by technology

In order to maintain the topological indicators such as node out-degree, clustering coefficient and average path length of the complete network, the strategy of randomly sampling some nodes and some edges will inevitably lead to the absence of important nodes or edges, and the topological data network formed after sampling It cannot form a connected network. When applied to a visualization scene, multiple connected subgraphs may be displayed, resulting in the inability to grasp the overall structure of the complete network and losing the meaning of topological network data sampling and visualization.

Images