Construction and application value of the grading prediction model of pressure injury in critically ill patients based on support vector machine algorithm
Construction and application value of the grading prediction model of pressure injury in critically ill patients based on support vector machine algorithm
Objective To construct and verify the grading prediction model of pressure injury(PI)in in-tensive care unit(ICU)patients based on support vector machine(SVM)algorithm.Methods The clinical da-ta of 157 inpatients in the Department of Critical Care Medicine of a tertiary hospital in Chongqing from De-cember 2020 to December 2022 were collected.The influencing factors of PI grading were screened by the Chi-square test and kruskal-Wallis H test.Then,the data were randomly divided into the training group and the validation group at a ratio of 7∶3.Based on the training group data,the support vector machine algorithm was used to establish the PI grading prediction model of ICU patients,and the five-fold cross-validation method was used to optimize the parameters.The trained model was internally validated in the validation group data set,and the confusion matrix analysis was performed on the results before and after.The performance of the model was evaluated by the accuracy rate,precision rate,recall rate,F1 value and the area under the curve(AUC)of the receiver operating characteristic value.Results The 10 factors affecting PI grading were pre-liminarily determined.The performance of the model was the best when gamma=0.1 and cost=2.2.The ac-curacy rate of the PI grading prediction model was 81.25％,the accuracy rate was 79.70％,the recall rate was 80.30％,and the F1 value was 79.90％.The AUC of the receiver operating characteristic was 0.939.Conclusion The constructed PI grading prediction model has good predictive performance,which can provide reference for clinical medical staff to formulate grading nursing intervention programs for preventing PI in ICU patients.
关键词
压力性损伤/支持向量机/高频超声/分级预测模型
Key words
Pressure injury/Support vector machine/High frequency ultrasound/Grading pre-diction model
维普
万方数据