BP神经网络在急性低血压发生预测模型中的应用

背景：ICU监护中术后急性低血压并发症的发生严重威胁着患者的生命安全,目前临床上主要依靠医生的经验进行预见性判断。 目的：为实现急性低血压发生的自动检测和提前预报,运用医学信息学理论,探讨一种预测急性低血压发生的模型。 方法：对发生与未发生急性低血压两者间平均动脉压信号进行小波多尺度分解,并选取各层小波系数的统计特征值中位数和最大值作为信号特征参数,提出了基于BP神经网络方法对提取的信号特征参数进行分类预测,并在MATLAB环境下进行仿真实验。 结果与结论：实验结果表明,利用BP神经网络方法对急性低血压发生的预测是可行的。

Application of back propagation neural networks in the prediction of acute hypotension

BACKGROUND:The post-operation complications of acute hypotensive episode (AHE) in intensive care units seriously endanger the patient’s lives, and it is depended mainly on the expert experience of doctors to treat. OBJECTIVE:To detect automatically and forecast the occurrence of AHE and to research a model for predicting AHE by medical informatics theory. METHODS:Mean arterial pressure (MAP) signals of those people who experienced AHE and those who did not experienced AHE were both described on different scales by using wavelet transform, and the median and maximum from the wavelet coefficients were extracted as the parameters of MAP signal. Then back propagation (BP) neural networks method for classifying and predicting the parameters was developed and simulated in MATLAB environment. RESULTS AND CONCLUSION:The experiment demonstrates that BP neural networks method was practicable for forecasting the occurrence of AHE.