Cerebral infarction probability prediction device

Abstract

The invention relates to a cerebral infarction probability prediction device, and the device comprises a data collection module which is used for collecting the related information of cardiovascular and cerebrovascular diseases of a patient; the probability prediction module is used for taking the doctor seeing information of the patient as the input of a probability prediction model to obtain the probability of cerebral infarction of the patient in the future; wherein the probability prediction model is constructed by a model construction module, and a model construction unit of the model construction module adjusts the logic regression model in a mode of adding a logic function; the training unit is used for adjusting the forms of the improved logistic regression models by setting different K values, and training the improved logistic regression models in different forms by adopting the cardiovascular and cerebrovascular disease related information; and the verification unit is used for verifying the trained logistic regression models in different forms in a cross validation mode, and taking the form of the logistic regression model corresponding to the proper K value as the probability prediction model. According to the invention, the prediction accuracy can be improved.

Description

technical field [0001] The invention relates to the technical field of medical big data in artificial intelligence, in particular to a cerebral infarction probability prediction device. Background technique [0002] With the improvement of people's living standards, more and more people begin to pay attention to their physical health status. Assessment of health status is becoming an increasingly common scene requirement in life, such as medical examination report analysis, health education and disease prevention, etc. . [0003] At present, the existing disease assessment models mainly use logistic regression to achieve classification, but for linear regression models with only 0 or 1 variables, they are relatively simple, and simple models have poor learning ability, so they are prone to underfitting The situation occurs, which leads to bias and makes it difficult to make accurate predictions. Contents of the invention [0004] The technical problem to be solved by the...

AI Technical Summary

Problems solved by technology

[0003] At present, the existing disease assessment models mainly use logistic regression to achieve classification, but for linear regression models with only 0 or 1 variables, they are relatively simple, and simple models have poor learning ability, so they are prone to underfitting The situation occurs, which leads to bias and makes it difficult to make accurate predictions

Images