Public opinion event prediction method and device based on social network and storage medium

Abstract

The invention provides a public opinion event prediction method and a device based on a social network and a storage medium, and the method can effectively improve the prediction precision of public opinion events, and comprises the following steps: collecting related social network data for each public opinion event, extracting keywords from the social network data several days before the occurrence of the public opinion event, constructing a keyword set every day, and extracting a propagation relationship among the users; carrying out corpus preprocessing on the social network data to obtaina corpus, and expressing all keywords as word vectors; constructing a semantic association information graph and a propagation relation graph of the keywords every day; constructing a dynamic dual-graph convolutional network model based on a graph convolutional network, inputting the semantic association information graph and the propagation relationship graph of the keywords of each day in a plurality of previous days into the dynamic dual-graph convolutional network model, and inputting previous data at the same time until the feature vectors of the keywords of the previous day are calculated; and performing nonlinear transformation, and outputting a prediction result of the public opinion event.

Description

technical field [0001] The invention relates to the technical fields of public opinion analysis, data mining and deep learning, and in particular to a social network-based public opinion event prediction method, device and storage medium. Background technique [0002] With the popularization and application of computers and the in-depth development of information technology, social media and news websites such as Weibo, forums, post bars, and blogs have become important battlefields for information dissemination. Netizens receive all kinds of information from the virtual world and the real world every day, and express their views and opinions on social events freely on social networks. When a network event breaks out in the virtual world or a social event breaks out in the real world, it may cause widespread concern and discussion among netizens, and then it will evolve into a public opinion event, which will have a significant impact on public security in society. At the s...

AI Technical Summary

Problems solved by technology

[0003] Traditional event prediction methods mainly use regression models, Bayesian networks, gray prediction models, etc. These methods have low prediction accuracy and have certain limitations, that is, they only consider public opinion events in one or two dimensions characteristics, because public opinion events have certain regularities in multiple dimensions such as users, content (semantics), time, and space (structure).

In order to utilize more features of public opinion events, event prediction methods based on deep learning are gradually emerging, such as early BP neural network, multi-instance learning (Multi-Instance Learning, MIL), recurrent neural network (Recurrent Neural Network, RNN), etc., these Deep learning methods have higher prediction accuracy than traditional methods, but the results are still not ideal

Images