多种统计模型在冠心病医疗费用影响因素中的应用与比较

发布时间：2019-03-17 07:46

【摘要】：目的:利用回归、人工神经网络、支持向量机建模,对医院住院费用的主要影响因素进行分析,寻找最有效的医疗费用影响因素分析模型。方法:以辽宁省10家县医疗机构2013年12月至2014年12月期间冠心病患者的住院信息为样本,建立支持向量机模型、BP神经网络模型、多元线性回归模型,分别对住院费用影响因素及其影响程度进行分析。结果:基于RBF核函数的支持向量机模型对住院费用预测准确度最高,达到87.39%。主要影响因素为住院天数、冠心病类型、是否有合并症、医疗机构、入院途径。结论:此次研究中,支持向量机在住院费用影响因素分析中具有很大的优势,建议在对数据进行分析时,可以采用多种模型进行分析,根据数据特点选取最合适的模型。
[Abstract]:Aim: to use regression, artificial neural network and support vector machine (SVM) to model the main influencing factors of hospital expenses, and to find out the most effective model of influencing factors of medical expenses. Methods: a support vector machine (SVM) model, a BP neural network model and a multiple linear regression model were established based on the inpatient information of coronary heart disease patients from December 2013 to December 2014 in 10 county medical institutions in Liaoning Province. The influencing factors and degree of hospitalization expenses were analyzed. Results: the support vector machine model based on RBF kernel function has the highest accuracy of 87.39%. The main influencing factors were length of hospitalization, type of coronary heart disease, complication, medical institution, admission route. Conclusion: in this study, support vector machine has great advantages in the analysis of influencing factors of hospitalization expenses. It is suggested that multiple models can be used to analyze the data, and the most suitable model should be selected according to the characteristics of the data.
【作者单位】：大连医科大学公共卫生学院
【基金】：辽宁省社会科学规划基金“大数据时代我省医保基金付费方式与监管策略研究”(编号:L15BGL008)
【分类号】：R541.4;R197.323

【参考文献】