Patient death risk prediction method and system based on electronic medical record, terminal and readable storage medium

Abstract

The invention discloses a patient death risk prediction method and system based on an electronic medical record, a terminal and a readable storage medium. According to the patient death risk prediction model constructed by the method, the time information is added to the medical feature data to obtain the patient feature data with the time characteristic, then the time sequence model is utilized to learn feature representation of various types of patient feature data, an irregular time pattern is mined, and the patient death risk prediction accuracy is improved. According to the method, the feature data of a patient is extracted, then an attention mechanism with a hierarchical structure is introduced to fuse the heterogeneous feature data to obtain a more comprehensive fused feature, and finally, the fused representation of the patient is applied to death risk prediction, so that the model prediction precision and reliability are improved. According to the method, the irregular time sequence modeling problem and the multivariate heterogeneous data fusion problem in the prior art are effectively solved, the method and other methods are tested and compared on the same data set, and experimental results show that the method has good performance in the aspect of death risk prediction of critical patients.

Description

technical field [0001] The invention belongs to the field of artificial intelligence and medical applications, and in particular relates to a method, system, terminal and readable storage medium for predicting patient death risk based on electronic medical records. Background technique [0002] The Intensive Care Unit (ICU) usually concentrates the most important resources of the hospital, such as departments, medical equipment and personnel, with the purpose of providing high-quality and comprehensive diagnosis and treatment services for critically ill patients, so as to reduce the mortality of patients. The in-hospital mortality rate of critically ill patients is easily affected by hospital resources and patients' own conditions. Globally, the mortality rate of patients in intensive care units remains high, and the high cost of diagnosis and treatment has greatly increased the living burden of patients and their families. Realizing the prediction of the death risk of pati...

AI Technical Summary

Problems solved by technology

Another challenging problem is the fusion representation of multivariate heterogeneous data

Images