Crowd information clustering method based on hyper-element heuristic algorithm

Abstract

The invention provides a crowd information clustering method based on a hyper-element heuristic algorithm. The crowd information clustering method solves the problem that it is difficult to establish crowd clustering in a dynamic environment in the prior art. The crowd information clustering method comprises the following steps: establishing a social graph model according to a social network of a crowd; obtaining an adjacent matrix A and elements a_{ij} of the matrix according to the model; obtaining the grade of the node i; by analyzing and judging the grade of the node i, determining a strong sensory group and a weak sensory group, so that a clustering target of the crowd information is defined, and the clustering target comprises a clustering target NRA and a clustering target RC; and designing a hyper-element heuristic algorithm to carry out clustering processing on the clustering target. According to the crowd information clustering method, the clustering capability of the algorithm in the dynamic network environment of the crowd information can be effectively improved, and the clustering quality is improved.

Description

technical field [0001] The invention belongs to the technical field of information processing, and relates to a crowd information clustering method based on a super-heuristic algorithm. Background technique [0002] At present, all kinds of social software in our country continue to develop and be widely used, bringing a lot of convenience to the people's life and entertainment. However, the use of social networks to carry out group-style illegal and criminal activities is also more concealed and more destructive, causing hidden dangers to the safety of people's lives and property. Therefore, by clustering the crowd data information in the social network, the personnel structure of different niche groups and their relationship in the social big data network will be effectively mined. [0003] In the past ten years, various types of network clustering algorithms have emerged, among which the more famous algorithms include Girvan-Newman algorithm, fast greedy module optimizat...

AI Technical Summary

Problems solved by technology

Although these single-objective algorithms are very successful in the theory and application of crowd information clustering, there are still the following problems that have not been resolved: 1. Different algorithms will produce different clustering patterns for the same network, that is, the crowd environment will change; 2. Users need to set the number of people for clustering, and it is impossible to adaptively set different algorithms according to the characteristics of the crowd information, that is, the behavior of individual personnel in the group will change

However, in the actual network, these information are usually unpredictable and have dynamic characteristics, so how to establish the clustering information of the crowd in a dynamic environment is a difficult problem in the current clustering analysis of crowd information

Images