一种基于IABC-RF的ICU患者心衰死亡率预测模型

Prediction Model of Heart Failure Mortality in ICU Based on IABC-RF

每年有将近5%的重症加强护理病房(Intensive Care Unit,ICU)患者入院是与心力衰竭相关的,针对ICU监测技术、治疗设备先进以及医疗资源高度集中,但患者死亡率仍然较高这一现象,已有研究主要利用多种危重病评分系统,评估ICU患者严重程度并预测患者死亡率用以辅助治疗.随着电子病历的发展,为了获得更优的预测性能,本文将大量重症监护数据库利用起来,结合了随机森林和改进蜂群优化算法,并考虑到不同合并症会加剧心衰的死亡率,提出了一种有效的ICU患者心衰死亡率预测模型(IABC-RF),在真实ICU病患数据集进行实验,结果表明,该方法能够对ICU患者死亡率进行有效的预测,有助于医生迅速判断病情潜在的风险、降低医疗周期,最终服务患者.

Nearly 5%of ICU admissions are associated with heart failure each year.Even if the ICU has advanced monitoring technology,treatment equipment and high concentration of medical resources,the mortality rate of patients still remains high.Previous studies have proposed multiple critical scoring models to predict patient mortality.To achieve a better performance,an effective ICU patient heart failure mortality prediction model(IABC-RF)is proposed in this paper,combined with random forest and improved ABC optimization algorithm.The results show that the new method can effectively predict the mortality of ICU patients.The accurate predictions will help doctors determine the potential risks of the disease quickly,reduce the medical cycle,and serve the patients finally.