基于BP神经网络的呼吸系统疾病诊断仿真研究

Simulation study of respiratory disease diagnosis based on BP neural network

文章选取某三甲医院的320例症状相似而诊断不同的呼吸系统疾病病例,将其分成训练数据和测试数据各160例,研究建立基于BP神经网络模型的呼吸系统疾病诊断仿真系统,运用Matlab仿真工具进行实验测试,将仿真诊断结果与经验丰富的临床医生的诊断进行比较,两者诊断符合率达到94.6%。结果表明,基于BP神经网络模型的呼吸系统疾病诊断仿真系统,可以模拟临床医生对呼吸系统疾病作出正确诊断。

320 cases of respiratory disease which have similar symptom but different diagnosis are selected from a 3A grade hospital,and they are divided into learning group and testing group with 160 cases each.A simulation system for respiratory disease diagnosis is presented based on BP neural network model,and the system is experimentally evaluated by using Matlab simulation tool.By comparing the simulation results with experienced clinicians＇ diagnosis opinions,it is found that there are up to 94.6% same diagnosis.It shows that the proposed simulation system can simulate clinicians＇ respiratory disease diagnosis correctly.