Causal network learning method based on local Granger causal analysis

Abstract

The invention belongs to the field of data mining, and provides a causal network learning method based on local Granger causal analysis. The method comprises the following steps: preprocessing acquired data, and complementing missing data by adopting an average value interpolation method; and carrying out stability test and processing on the complemented data so as to meet the hypothesis of establishing a model; and then normalizing the data to eliminate the influence caused by different variable dimensions; and finally, building a causal network learning algorithm based on local Granger causal analysis, so that the purpose of accurately exploring the causal relationship between the variables is achieved, and meanwhile, a dynamic causal relationship curve between different variables is displayed, so that the causal relationship between the variables between systems is quantitatively and clearly analyzed.

Description

technical field [0001] The invention belongs to the technical field of data mining, relates to a causal network learning method based on local Granger causality analysis, and aims to explore the relationship between variables in high-dimensional data in fields such as meteorology. Background technique [0002] Multivariate time series is a set of discrete observations of multiple variables distributed according to time, which widely exist in many fields such as finance, industry, transportation, and meteorology. For example, in the field of air pollution research, in recent years, with the rapid development of my country's industry and the increase of vehicles, the concentration of harmful substances in the atmosphere due to the combustion of coal and oil has increased significantly, resulting in the decline of air quality and smog weather. happened. Haze weather is a state of air pollution, which is a general expression of the excessive content of various suspended particul...

AI Technical Summary

Problems solved by technology

[0006] The technical problem to be solved by the present invention is that the traditional Granger causality analysis method cannot be applied to high-dimensional time series due to structural limitations, and cannot explore the real-time dynamics between variables in the process. The analysis model has been expanded, and a causal network learning algorithm based on local Granger causality analysis is proposed to realize accurate causal relationship exploration of high-dimensional data and real-time display of information in the process

Images