Network coupling time sequence information flow prediction method based on causal logic and graph convolution feature extraction

Abstract

The invention discloses a network coupling time sequence information flow prediction method based on causal logic and graph convolution feature extraction. The method comprises the following steps: establishing a causal logic network by using transfer entropy based on node time sequence data, performing feature extraction on logic network node data by using a graph convolutional network (GCN), performing flow prediction by using a gated cycle unit (GRU) on the basis of a graph information feature h, and performing training optimization on parameters by using a back propagation algorithm. By adding the logic information of the causal network and combining the graph convolution network and the gating cycle unit, the information flow prediction precision is improved.

Description

technical field [0001] The invention relates to a network coupling time series information flow prediction method based on causal logic and graph convolution feature extraction, which belongs to the technical field of information flow prediction. Background technique [0002] Modern management services have entered a new era driven by data resources and information technology. Big data provides a broader perspective and a more effective approach for information management and services. Relying on the "data intelligence" formed by multi-source big data resources, it is urgent to build a development model guided by the big data system and driven by technological innovation to meet the needs of modern information management services. First of all, it is necessary to improve the networked and standardized data state perception collection system and further integrate data resources; second, it is necessary to innovate big data analysis applications to support efficient operation ...

AI Technical Summary

Problems solved by technology

[0006] The purpose of the present invention is to address the above-mentioned problems existing in the existing technology, and to provide a network coupling time-series information flow prediction method based on causal logic and graph convolution feature extraction. The technical problem to be solved by the present invention is to calculate the The transfer entropy among them establishes a causal logic network that describes its dependencies, and replaces the single-entity traffic network as the input of coupling time-series information flow prediction, so that the graph convolutional network (GCN) is used to extract the graph data on the basis of the causal logic network. features, and then use the extracted features to predict using the Gated Recurrent Unit (GRU)

Images