From-bottom-to-top high-dimension-data causal network learning method

Abstract

The invention discloses a from-bottom-to-top high-dimension-data causal network learning method. The method includes the steps of a causal relationship local structure discovery algorithm, wherein a local causal relation learning method and a causal relationship intensity communication strategy are adopted to learn the local causal relationship intensity relationship among variables; a global variable causal sorting algorithm, wherein on the basis of the biggest loop-free directed subgraph model, high-dimension variable global causal relationship sorting is achieved on the basis of local structure strength measurement and a redundant causal relationship elimination strategy, wherein on the basis of global causal sorting, reliable causal relationship discovery on high-dimension observation data is finally achieved.

Description

technical field

[0001] The invention relates to the field of data mining, in particular to a bottom-up causal network learning method oriented to high-dimensional observation data. Background technique

[0002] At present, causal inference has been widely used in various fields, typical applications such as biological network inference, disease diagnosis, drug effect analysis, disease-causing gene discovery, social network analysis, etc. The application requirements in these fields have prompted a lot of causal discovery research work, and a large number of causal inference theories and algorithms have emerged. The basis of causal inference theory, algorithm and application is the causal relationship model. The classic causal relationship models include the Rubin Causal Model (Rubin Causal Model; RCM) proposed by Donald Rubin and the Causal Diagram proposed by Judea Pearl. Pearl illustrates the equivalence of the two. The former (Rubin causality model) examines the averag...

Images