基于集成支持向量机的医疗设备风险评估研究与应用

Research and application of medical equipment risk assessment based on integrated support vector machine

医疗设备临床使用风险的精确评估是风险预警和防控的基础。本研究将德尔菲法与相关性分析、鉴别力分析相结合,从设备、环境、人因、管理4个方面建立指标体系,提出了基于支持向量机(support vector machine,SVM)的风险评估方法,运用自适应增强算法(adaptive boosting,AdaBoost)进一步提高算法的泛化性能和评估精度,构建医疗设备临床使用风险评估模型。将该模型用于三甲医院ICU在用呼吸机风险因素的综合评估,N折交叉验证模型的适用性,取得了良好的应用效果。

Accurate risk assessment of clinical medical equipment is the precondition of risk warning and controlling.In this study,the Delphi method was combined with correlation analysis and discrimination analysis to determine the evaluation index system from the four aspects of equipment,environment,human factors and management.A risk assessment method based on support vector machine(SVM)was proposed.AdaBoost algorithm was used to improve the generalization performance and evaluation accuracy of SVM.The model is applied to the comprehensive evaluation of ventilator risk factors in the ICU,and the applicability of the model is verified by N-fold cross validation.