基于BP神经网络的MRI检查总用时预测方法研究

Research on total time prediction method of MRI examination based on BP neural network

目的:提出一种基于BP神经网络的MRI检查总用时预测方法。方法:对某三甲医院MRI设备临床使用数据展开深度挖掘,将63种检查部位与26种检查方法组合为若干检查序列,并将检查序列向量作为BP神经网络的输入、患者MRI检查总用时作为输出,构建BP神经网络模型。以平均绝对误差(mean absolute error,MAE)作为评价指标,对比BP神经网络模型与岭回归模型、Lasso回归模型预测MRI检查总用时的性能。结果:BP神经网络模型的预测值与实测值更为接近,MAE为53.14 s,均优于岭回归模型、Lasso回归模型。结论:提出的方法能准确地预测患者的MRI检查总用时,能够更好地帮助医院工作人员安排患者检查。

Objective To propose a total time prediction method of MRI examination based on BP neural network.Methods The clinical data of MRI equipment in some tertiary grade A hospital was mined deeply.Totally 63 examination sites and 26 examination methods were combined into several examination sequences,and the examination sequence vector and total examination time were used as the input and output of BP neural network,respectively,so as to construct a BP neural network model.The BP neural network model established was compared with the ridge regression model and Lasso regression model with the mean absolute error(MAE)as the evaluation index when used for predicting MRI examinations.Results The BP neural network model behaved better than the ridge regression model and Lasso regression model with the predicted value close to the measured value and MAE being 53.14 s.Conclusion The proposed method accurately predicts the total MRI examination time and can assist hospital staff in scheduling patient examinations.