ICU death rate prediction method and system based on punishment integrated model

Abstract

The invention provides an ICU death rate prediction method and system based on a punishment integrated model. The ICU death rate prediction method based on the punishment integrated model comprises the following steps: acquiring multiple original data characteristics of an ICU patient from multiple dimensions; preprocessing the plurality of the acquired original data characteristics; mining and extracting new data characteristics on the basis of the original data characteristics; selecting the original data characteristics and the new data characteristics based on an algorithm in the integrated model to form an input characteristic set; and inputting the formed input characteristic set into a trained and tested integrated model to obtain an ICU death rate prediction result, wherein the integrated model integrates a logistic regression algorithm based on a weight punishment strategy and a LightGBM algorithm based on the weight punishment strategy.

Description

technical field [0001] The invention relates to the field of ICU mortality prediction, and in particular to an ICU mortality prediction method and system based on a penalty integrated model. Background technique [0002] The intensive care unit (ICU) concentrates the hospital's most advanced monitoring equipment and first aid facilities, which makes it play an important role in reducing the mortality rate. Predicting the discharge mortality rate of ICU patients helps hospitals rationally allocate medical resources on the one hand, and on the other hand helps clinicians formulate diagnosis and treatment plans, thereby reducing the mortality rate of ICU patients. However, the data formed by ICU equipment is rich and complex, and usually exhibits the characteristics of high dimensionality, imbalance and time asynchrony. Therefore, even experienced clinicians cannot quickly and accurately judge the progression of the disease or the extent to which the disease is affecting the p...

AI Technical Summary

Problems solved by technology

These scoring methods can help reduce the pressure on clinicians, but considering that they are all based on linear models, the mortality prediction performance is somewhat limited

In addition, there is a large difference between the number of survivors and deaths of ICU patients, and the traditional method does not consider the impact of the unbalanced data distribution caused by this difference on the scoring model

Images