基于BP神经网络的脑血管疾病院内感染风险预测模型研究

发布时间：2018-06-08 22:03

本文选题：脑血管疾病 + 院内感染　；参考：《重庆医学》2017年12期

【摘要】：目的采用BP神经网络构建脑血管疾病院内感染预测模型,对患者发生院内感染的风险作出评判,加强脑血管疾病院内感染防控管理。方法收集整理桂林市中医医院2014年脑病科892例患者院内感染数据,采用主成分分析法选取具有代表性的指标并予以赋值,依指标集构建基于BP神经网络的脑血管疾病院内感染风险评估模型,通过一定数量样本的训练和测试,以检验模型可靠性和准确性。结果选取200例患者作为样本对神经网络进行测试,未发生院内感染病例预测准确174例,准确率89.2%;发生院内感染病例预测准确5例,准确率100%。结论 BP神经网络风险评估模型对预防脑血管疾病院内感染有着积极的意义,可较好的提前预测院内感染发生风险,为感染防控提供有效的依据。
[Abstract]:Objective to establish a prediction model of nosocomial infection in cerebrovascular diseases by using BP neural network to evaluate the risk of nosocomial infection and strengthen the prevention and control management of nosocomial infection of cerebrovascular diseases. Methods the nosocomial infection data of 892 patients in the Department of Encephalopathy in Guilin traditional Chinese Medicine Hospital in 2014 were collected, and the representative indexes were selected and assigned by principal component analysis (PCA). A risk assessment model for cerebrovascular diseases in hospital was constructed based on BP neural network. The reliability and accuracy of the model were tested by training and testing of a certain number of samples. Results 200 patients were selected as samples to test the neural network, 174 cases of nosocomial infection were predicted accurately, and 5 cases of nosocomial infection were predicted accurately, and 100 cases of nosocomial infection were predicted. Conclusion BP neural network risk assessment model has positive significance in preventing nosocomial infection of cerebrovascular diseases, which can predict the risk of nosocomial infection in advance and provide effective basis for prevention and control of infection.
【作者单位】：桂林市中医医院医院感染管理科;重庆市中医院消化内科;
【分类号】：R743

【参考文献】