基于时间序列模型的医保记账费用预测研究

Study on Forecasting of Medical Insurance Accounting Cost Based on Time Series Model

目的:探讨时间序列模型在医保记账费用预测中的可行性,并预测医保记账费用的发展趋势。方法:收集江苏省某三甲医院2012—2016年每月医保记账费用的原始数据,拟合时间序列模型,并对模型的拟合程度和预测效果进行检验。结果:模型通过白噪声检验,模型的平均绝对百分误差(MAPE)是5.979%,均方根误差(RMSE)是5 052 736.327,平均绝对误差(MAE)是3 612 828.978,R方为0.912,模型的拟合度良好。结论:将此模型用于2017年医保记账费用预测,模型的拟合值与实际值拟合程度较好,能为医院在医保记账费用管理中的决策提供参考依据。

Objective:to explore the feasibility of the time series model in the prediction of the medical insurance accounting and to predict the development trend of the medical insurance accounting.Methods:to collect the original data of the monthly charge of the medical insurance in a Sanjia hospital of Jiangsu province from 2012 to 2016,applying into the time series model,and to test the fitting degree and prediction effect of the model.Results:According to the white noise test,the model of the mean absolute percentage error(MAPE)is 5.979%,the root mean square error(RMSE)is 5052736.327,the mean absolute error(MAE)is 3612828.978,0.912 R party,meeting the expectation.Conclusions:this model was used to forecast the medical care accounting in 2017,and the forecasted indexes fit the real indexes well.It shows that the model can provida reference for decision making in medical insurance management.