Social network influence maximization method based on activeness

Abstract

The invention discloses a social network influence maximization method based on activeness. The social network influence maximization method comprises the following steps: constructing a simple network G'; selecting a node with a high activeness ACT attribute value according to a node activeness ACT attribute to enter an ACT node set H; calculating a node influence synthetic value aps (v) according to the sizes of the node activeness ACT attribute value and the influence; circularly selecting k seed nodes; updating the node influence synthetic value aps (v) in a network when each seed node is selected. A result shows that the activeness attribute is introduced based on an IC model; the extension of the model enables a social network influence maximization problem to be more scientific; an ACH (Automated Clearing House) algorithm is close to an influence range of a KK greedy algorithm in an influence range and the social network influence maximization method has very good representation on timeliness.

Description

technical field [0001] The invention belongs to the application field of computer information technology, relates to a social network influence maximization technology, in particular to a method for selecting seed nodes based on activity degree to maximize social network influence. Background technique [0002] Social networks are composed of a large number of users and complex relationships among users (including family relationships, friend relationships, classmate relationships, work relationships, etc.). Different from traditional networks, the dissemination and diffusion of information in social networks depends on the relationship between users. How to make information be received by as many users as possible in the network, that is, the problem of maximizing the influence of social networks, is a current research hotspot in social networks and their applications. Richardson, M. in the literature "Mining knowledge-sharing sites for viral marketing" and Domingos, P. in ...

AI Technical Summary

Problems solved by technology

[0004] However, it can be seen from the actual analysis that the selection effect of the initial node depends on two aspects, firstly, how to define the influence of the node, and secondly, the effect of algorithm execution also has a great influence

The independent cascading (IC) information propagation model commonly used in existing research generally calculates the influence of nodes based on the number of neighbor nodes and the activation probability of nodes to neighbor nodes. It is not difficult to analyze. There are two problems in this calculation method : The first one lies in influential users, that is, some users with high influence. Due to their low activity, they do not really exert their influence to transmit information. These influential nodes should not be selected into the seed node set ; The second problem lies in the activation of users with limited transmission power. In the traditional IC model, it is believed that the node in the active state will perform reactivation behavior on its neighbor nodes 100%, but the fact is that the reactivation power of the activation user is much lower than 100%. , which is related to when the whole propagation process ends

Images