人工神经网络在冠心病中的仿真研究

Simulation on Coronary Heart Disease Based on Neural Network

研究优化冠心病无伤害准确诊断,冠状动脉造影和心电图是冠心病诊断的主要方法,但冠状动脉造影诊断是一种创伤性介入疗法,会对患者身体造成一定的损伤,心电图方法存在着较大的误诊率。根据冠心病与一些常规的指标有一定的联系,提出BP神经网络的冠心病诊断模型,实现冠心病的无损伤性诊断。首先对各诊断指标进行量化,然后构建BP神经网络诊断模型,利用样本数据集进行训练,获得最优的冠心病诊断模型,最后利用测试集对模型进行验证。仿真结果表明,利用BP神经网络进行冠心病诊断,诊断的准确率高,达到了冠状动脉造影的诊断效果,克服了传统的诊断方式对身体造成损伤或误诊率高的缺限,是一种高效的冠心病辅助诊断方法。

Coronary autobiography and electro cardiogram methods are main diagnosis methods in coronary heart disease,but the coronary angiography diagnosis method is a traumatic intervention therapy,which can cause certain damage in patients,and electro cardiogram methods has high misdiagnosis rate.Because some general index have a link with coronary heart disease,the paper puts forward the BP neural network of coronary heart disease diagnosis model,which can realize noninvasive diagnosis of coronary heart disease.Firstly,quantifications all diagnosis index,and then construct the BP neural network model,training the BP neural network with the diagnostic data sets,we obtain the optimal coronary heart disease diagnosis model,lastly,we use the test set to verify diagnosis model.Simulation results show that the BP neural network has high diagnosis accuracy rate of coronary heart disease,reaches the effect of coronary angiography,and overcomes the defects of traditional diagnosis way which damages the body or has high misdiagnosis rate.The result shows that the BP neural network is a high efficient auxiliary coronary heart disease diagnosis method.