基于双谱分析—BP神经网络的ICU急性低血压发生预测模型研究

Study on Predicting Model for Acute Hypotensive Episodes in ICU Based on Bispectrum Analysis and BP Neural Network

目的:ICU中,急性低血压的发生严重威胁着患者的生命安全,能够及时准确地对其预测具有重要临床意义。为了提高医生对患者发病提前判断的准确性,本文研究了一种基于双谱分析和BP神经网络的急性低血压发生的预测模型。方法:应用双谱分析提取动脉血压数据特征,构建特征向量,利用BP神经网络训练出分类预测模型,实现对急性低血压发生的提前预测。结果:经过不断的优化调整,本文最终构造了一个三层的BP神经网络预测模型,具有良好的自动预测能力。结论:实验表明,本方法能达到比较好的分类预测效果,可为ICU中急性低血压发生的提前预测和干预提供辅助参考。

Objective： The occurrence of acute hypotensive episodes（AHE） in intensive care units（ICU） seriously endanger the lives of patients.Its timely and accurate prediction has important clinical significance.To help the doctors improving the accuracy of AHE early judgment,a model based on bispectrum analysis and BP neural network is proposed in this paper. Methods： We use bispectrum analysis to extract features of arterial blood pressure（ABP） signal and construct feature vector and train BP neural network to develop a predicting model to predict AHE in advance.Results： After continuous optimization and comparison,a three-layer BP neural network is constructed in this paper.This neural network is proved to have good automatic predicting ability.Conclusions： Experiment demonstrates that this model performs well on prediction and classification,which will be conductive to provide auxiliary reference for early AHE predicting and intervention in ICU.