A Microblog Popularity Prediction Method Based on Time and User's Forwarding Sequence

Abstract

The present invention relates to a microblog popularity prediction model based on time and user's reposting sequence, which belongs to the field of message popularity prediction in social networks, and includes the following steps: S1: using a recurrent neural network to model the reposting sequence of microblog , used to capture the long-distance dependence of the message propagation process; S2: The output of the hidden layer is transformed into a nonlinear network to learn the rate of each time step in the propagation process; S3: The early trend acceleration obtained by using the rate and the early Popularity, and under the optimization of user activity, predict the future popularity of Weibo. The invention ensures more accurate prediction of the future popular trend in the early stage of news dissemination, and the model not only utilizes the historical dissemination information, but also well describes the microblog dissemination process.

Description

technical field [0001] The invention belongs to the field of message popularity prediction in social networks, and relates to a microblog popularity prediction model based on time and user forwarding sequence5 Background technique [0002] The ubiquity and cheapness of Web 2.0 services has changed the way content is generated and consumed online. In recent years, Internet technology has developed rapidly. With the rapid rise and popularization of the Internet, our lives are now inseparable from the Internet. Due to the Internet, content producers can use traditional channels to reach audiences of unimaginable scale, and realize the services of communication between content producers and consumers on a global scale, including video, photo, music sharing, weblogs, social bookmarking sites, Collaborative portal and content submission, browsing, content news aggregator rating and discussion etc. Social networking services represented by Facebook, Twitter, Weibo, WeChat, etc. h...

AI Technical Summary

Problems solved by technology

Models based on classification or regression rely on the extraction of features and do not describe the process of message dissemination. The performance of methods based on point random processes is lacking, and due to the diversification of social networks, they cannot be adapted to every social network. And there is no supervision using historical information

Images