中文摘要：

目的：构建基于人工神经网络的公立医院运行监管指标体系,为完善公立医院运行监管提供参考。方法：定性与定量相结合,现场调研与数理模型相结合。通过文献分析与专家咨询确定指标,收集118家医院指标数据,构建神经网络模型,计算指标权重。结果：确定了药品收入占业务收入的比例、负债率、医生护士人数比等7项指标及其权重,利用该指标体系得到测算值和实际值之间无显著差异。结论：本研究所得的指标体系具有较高的信度和效度。

英文摘要：

Objective： To construct the supervision index system operated in public hospitals based on artificial neural network, provide references for implementing the supervision of public hospital operation. Methods： To integrate qualitative analysis and quantitative analysis, integrate site investigation and mathematical model. Through literature analysis and decide index by expert consultation, collect index data of 118 hospitals, construct the neural network model and calculate the index weight. Results： 7 indexes, such as proportion of medicine income in business income, debt ratio, doctors and nurses ratio and the weights of these indexes are concerned. There is no significant difference between the measure value and actual value by using this index system. Conclusion： The index system given by this analysis has higher reliability and validity.