Group oriented influence maximizing method

Abstract

The invention discloses a group oriented influence maximizing method. The method comprises the following steps: first collecting and settling historic infection data of a group, and regarding the points in the same group as random variables with the same probability (namely homogeneity assumption), computing a complete probability space D of a point set X under the homogeneity assumption, independently describing the structured association of the group through the condition probability of the group on the historic data, and constructing an association graph IG accoridng to the entropy computation, and then selecting a seed set S with the maximum influence range size as k through the influence range computation on the association graph by use of a greedy algorithm. Through the adoption of the method disclosed by the invention, the most powerful group seed set can be rapidly positioned independent of the acquisition of the point influence relation; and the method is more effective and accurate in comparison with the ordinary algorithm when the group number in the network is far less than the point number.

Description

technical field [0001] The invention belongs to the field of information dissemination, in particular to a group-oriented influence maximization method. Background technique [0002] The rapid development of the Internet has led to the formation of an observable social network, which provides great convenience for the study of phenomena such as information dissemination and disease spread. The problem of finding a limited subset in a given network and maximizing the influence of the subset according to the cascade transfer of influence has received extensive attention. At present, the research object of influence maximization mainly focuses on entity points (such as individuals or blogs), and through the influence relationship of these entity points, related algorithms are designed to search for k-point combinations with the greatest influence. However, in real life, people tend to analyze the influence of the combination of groups (such as communities or various groups of ...

AI Technical Summary

Problems solved by technology

The inter-group influence is essentially the influence of inter-group points, but the influence maximization analysis based on the group granularity leads to the invisible point influence relationship, which makes the inter-group influence uncertain. At the same time, the group as a collection of points It may be affected by multiple neighbors at the same time and the state is a continuous value, making it necessary to establish more complex rules to calculate the impact size when dynamically simulating group influence transfer

Images