Self-adaptive information spreading source detection method

Abstract

The invention discloses a self-adaptive information spreading source detection method which can be applied to social network public opinion control. The problem that in a social network, due to the fact that information types, an unreal information spreading mode and an information exchange mode between users are complex and changeable, the task of detecting a spreading source of information is very difficult is solved. According to the method, firstly, a learnable logical relation graph is constructed in consideration of various information communication modes such as friends, forwarding and comments among users so as to adapt to a complex relation among the users; the invention also designs an unreal information centrality measurement method which can be coupled into a deep learning model for parameter learning so as to adaptively calculate the network centrality of each node in the unreal information propagation network. In order to flexibly process the noise in the social network, the invention further provides a noise elimination mechanism based on clustering.

Description

technical field [0001] The invention belongs to the field of social networks, and in particular relates to an adaptive information dissemination source detection method. Background technique [0002] Social platforms allow users to share and disseminate content more freely, which is very attractive to users. In addition, a variety of colorful pictures, texts, videos and other information also make people more and more inseparable from social media in their daily lives. Based on the huge number of users, the exposure and dissemination of information is also higher. As an information dissemination platform, its efficient, fast and broadcast characteristics also provide conditions for the generation and dissemination of false information, and social platforms such as Weibo and Twitter often lack a qualified audit mechanism to judge the authenticity of the information posted by users. The convenience of social networks means that the cost and speed of fabricating and dissemina...

AI Technical Summary

Problems solved by technology

[0006] (2) The method of multi-source detection: Zang W, Zhang P, Zhou C, et al. Discovering multiplediffusion source nodes in social networks [J]. Procedia Computer Science, 2014, 29: 443-452. Using community clustering algorithm to combine multiple The source localization problem is transformed into multiple single source localization problems, and then the node with the maximum likelihood estimation is identified as the source node on the infected cluster, but this method is only for the SIR model

The current method of detecting the source of information dissemination mainly assumes that false information is spread through the friendship of users. In fact, the way of spreading false information among users is not only simple friendship, forwarding and commenting will cause false information to spread. spread further

[0010] (3) Although it is intuitive to use network centrality to measure the ability of nodes to disseminate false information, on the one hand, it is difficult to predict which centrality measurement method should be selected; With huge computing power, most of the existing centrality measurement methods in large-scale real social networks are not advisable

[0011] (4) The social network itself is complex, and it is difficult for us to select the batch of nodes that may be the source from the many nodes. Existing methods pay little attention to how to eliminate the noise in the social network that has nothing to do with false information detection

Images