中文摘要：

目的：建立基于BP神经网络的治疗结果拟合模型，并在已建立的神经网络模型的基础上，进行中医治疗结果的预测和影响因素的敏感度分析，利用本研究的建模结果，为BP神经网络建模的方法学提供一定的参考依据，并能帮助医务人员做出正确的决策和分析。方法：在SPSSClementinel2．0中进行建模和预测，预测结果用SPSS13．0进行ROC分析。结果：BP神经网络的拟合度和预测准确度为81．224％，门静脉内径、中医症候积分对患者的治疗结果影响最大。结论：BP神经网络是比较适合于治疗结果数据特征的建模方法。

英文摘要：

Objective： To establish the sensitivity of treatment outcomes based on BP neural network fitting model, and in the established neural network model based on the prediction of the results and impact factors of the Chinese medicine treatment analysis, the use of the modeling results of this study school to provide some reference for the BP neural network modeling method, and can help the medical staff to make the fight decisions and analysis. Methods： SPSS Clementinel2.0 modeling and forecasting, and forecast results SPSS13.0 ROC analysis. Results： BP neural network goodness of fit and prediction accuracy for 81.224%, portal vein diameter, TCM syndrome scores the greatest impact on patient outcomes. Conclusion： BP neural network is suitable for the treatment of the characteristics of the resulting data modeling method.