Time series community and topic detection method based on weighted time series text network

Abstract

The invention discloses a time-series community and topic detection method based on a weighted time-series text network. The Booth sampling method constructs the inference process of the generative model; according to the inference process of the model, the weighted time series text network is trained to extract the community information, topic information, the correspondence between the community and the topic, the user's influence and participation in the community Time-varying characteristics; predict user behavior based on extracted information. The invention carries out a new modeling for the time information and weight information in the time series text network, considers the time information of the edges in the network and carries out continuity modeling, and comprehensively models the weighted time series network, which is beneficial to Understand the changes and development of communities on time scales and individuals about the development of communities on time scales.

Description

technical field [0001] The invention relates to the field of text network exploratory search, in particular to a time series community and topic detection method based on weighted time series text network. Background technique [0002] With the advent of the era of big data, it means that the amount of global data has grown exponentially. As one of the sources of data volume, online social media, whether blogs, video sharing sites, or social networks, has experienced rapid growth in the past half a decade. Faced with such a large amount of data, it is necessary to extract meaningful information from it, and the internal network structure of these data is a very important basis for extraction. So learn as much as possible about the social network structure. One of these methods is to identify groups of nodes with the same properties or functions, which is known as "community discovery". [0003] For weighted dynamic text networks such as online social media tweets and acad...

AI Technical Summary

Problems solved by technology

Another problem is how to use dynamic information to segment on the time axis to obtain time-divided sub-networks. Because the distribution of the network on the time axis is continuous, how to allocate the time for segmentation is a thorny issue.

Images