Model training method for supporting precise prevention, diagnosis and treatment of congenital heart disease

Abstract

The invention discloses a model training method for supporting precise prevention, diagnosis and treatment of a congenital heart disease, and the method comprises the steps: obtaining the index data of a patient, and determining an independent variable index corresponding to each dimension in the index data of the patient based on a preset rule; determining a value rate of the independent variable index of each dimension, determining a plurality of target independent variable indexes in each independent variable index based on the value rate, and determining a plurality of independent variable index groups based on each target independent variable index; respectively inputting each independent variable index group into each preset model corresponding to the to-be-predicted event to carry out model training so as to obtain a plurality of prediction models; and determining the directivity goodness of fit of a plurality of prediction models corresponding to each independent variable index group, and determining a target model corresponding to the to-be-predicted event based on the directivity goodness of fit. According to the method, the independent variable indexes are screened through the value rate, the defect indexes are eliminated in the model training process, and the accuracy of model training is improved.

Description

technical field [0001] The invention relates to the technical field of data processing, in particular to a model training method supporting precise prevention, diagnosis and treatment of congenital heart disease. Background technique [0002] With the vigorous development of medical informatization in recent years, medical big data has grown explosively. Based on medical data and machine learning technology, learning doctors' diagnosis and treatment behaviors, giving diagnosis and treatment suggestions, and assisting junior doctors to make clinical decisions have always been the hot direction of medical artificial intelligence. [0003] Existing technologies are generally based on traditional machine learning methods such as Bayesian, logistic regression, decision tree and svm for disease diagnosis and prediction. With the rise of neural networks in recent years, there are also some works based on deep neural networks for modeling. [0004] But in the prior art, although i...

AI Technical Summary

Problems solved by technology

[0006] The main purpose of the present invention is to provide a model training method that supports accurate prevention, diagnosis and treatment of congenital heart disease, aiming to solve the technical problem of inaccurate models obtained by model training on existing medical data

Images