Method for ordering microblog network node weights based on multi-link

Abstract

The invention discloses a method for ordering microblog network node weights based on multi-link, which is used for solving the technical problem of low recall rate in existing methods for ordering microblog network node weights. The technical proposal is as follows: a microblog network is constructed as a network directed graph G = (E, V), and the scale of the network graph is then remarkably reduced by defining a user effective fan set Ef(u) to improve the calculation speed; secondly, a node message is tracked, the issuing, forwarding, replying and collecting of the node message and other information are recorded, the additional link relation of nodes is determined, and the content of specific messages is not analyzed in the section so as to enhance the network analysis speed; and finally, the interaction of network nodes is identified and processed in combination with network dynamics, thereby lowering overall system overhead and improving the accurate rate and recall rate of the calculation method.

Description

technical field [0001] The invention relates to a method for sorting the weights of microblog network nodes, in particular to a method for sorting the weights of microblog network nodes based on multiple links. Background technique [0002] Micro-Blogging Network (Micro-Blogging Network) provides users with an interactive communication platform, which enriches users' spare time life and enhances communication among friends. [0003] In traditional social networks, node weights are mainly calculated by calculating the number of user friends and analyzing the structural relationship between users. Usually, the node degree in the network structure is analyzed, and the accuracy rate is not high simply considering the out-degree. In the microblog network, the network nodes constitute a directed graph, and only considering the out-degree and in-degree (In-Degree) cannot achieve high accuracy. [0004] Document 1 "Discovering Important Bloggers based on Analyzing Blog Threads [WW...

AI Technical Summary

Problems solved by technology

[0004] Document 1 "Discovering Important Bloggers based on Analyzing Blog Threads [WWW, Chiba, Japan, May 10-14, 2005]" discloses an important blog user analysis method ThreadRank based on post content analysis, which analyzes a large number of blog content To judge the importance of its users, it takes a lot of time for content cleaning and analysis, and the efficiency is low

[0005] Document 2 "Identifying Opinion Leaders in the Blogosphere [CIKM, pp.971-974, 2007]" discloses an opinion leader identification method, InfluenceRank, which judges the importance of users based on comparison with other blogs, and the importance of these users to blogs. The contribution made by the entire network is used to calculate the user weight. This paper uses the cosine theorem to calculate the similarity of different blog entities, which has high complexity and high overhead.

[0006] Document 3 "TwitterRank: Finding topic-sensitive Inuential Twitterers [WSDM, 2010]" discloses a TwitterRank calculation method for Twitter network nodes. The role played by various user groups in the process is used for weight calculation. This algorithm is mainly based on topic analysis, and the recall rate is not high.

Images