Diabetes risk factor cause and effect discovery method based on improved function cause and effect likelihood

Abstract

The invention discloses a diabetes risk factor cause and effect discovery method based on improved function cause and effect likelihood, and belongs to the technical field of medical informatization.The method comprises the steps: obtaining joint distribution of diabetes risk factor variable subsets; solving the log likelihood of the observation data according to the joint distribution and the cause and effect structure; converting the logarithm likelihood of the observation data into logarithm likelihood of noise of the observation data, and then establishing a diabetes risk factor FCL model; and correcting the diabetes risk factor FCL model by adjusting a threshold value to obtain a diabetes risk factor IFCL model, and discovering the cause and effect relationship of risk factors by utilizing the diabetes risk factor IFCL model. According to the method, the adjustment threshold value is introduced, the diabetes risk factor IFCL model is constructed, the cause and effect relationshipof the risk factors is found by utilizing the diabetes risk factor IFCL model, redundant edges and wrong edges of a diabetes risk factor cause and effect structure are reduced, and then the optimizeddiabetes risk factor cause and effect structure is generated.

Description

technical field [0001] The invention relates to the technical field of medical information technology, in particular to a causal discovery method for diabetes risk factors based on improved function causal likelihood. Background technique [0002] The number of diabetic patients is increasing year by year, and it has become the third major disease that threatens human health after cardiovascular and cerebrovascular diseases and malignant tumors. Analyzing the relationship between various risk factors and the relationship between risk factors and diabetes is the key to revealing the pathogenesis and pathology of diabetes, and it is also the premise of diabetes prevention and treatment. [0003] At present, research at home and abroad mainly focuses on the analysis of diabetes risk factors and the construction of diabetes prediction models. The research on risk factors analysis of diabetes mainly includes two aspects: the discovery of new risk factors and the correlation anal...

AI Technical Summary

Problems solved by technology

[0004] Randomized controlled experiments are the traditional method for discovering causality, but they require extensive intervention in the experimental group, which is not only costly, but also potentially unethical

The causality discovery method based on observation data can avoid the above problems, but the noise in the data will affect the causality discovery effect

Under the condition of significant noise, the causal relationship can be found effectively based on the FCL algorithm [Document: Ruichu Cai, Jie Qiao, Zhenjie Zhang, Zhifeng Hao.SELF: Structural Equational Embedded Likelihood Framework for Causal Discovery.AAAI.2018. On the problem of causality discovery of risk factors, more redundant edges and error edges will be retained

Images