基于灰色马尔柯夫模型的医院出院量预测

The Forecast of the Hospital Discharged Number Based on The Grey-Markov Model

目的结合GM(1,1)直接建模法和马尔柯夫预测原理,构建灰色马尔柯夫预测模型,预测出院量人次,为医院的管理决策提供统计学依据。方法根据福建某医院2005年-2015年的出院量数据,应用MATLAB软件,通过直接建模法建立出院量的GM(1,1)模型,结合马尔柯夫预测原理,对原始数据进行3个状态区间的划分,并由转移概率矩阵确定未来状态,从而得到最终预测值。结果 2013年-2015年预测值的相对误差分别为0.05%,0.35%和0.14%,2016年出院量预测值为108 335人次,增长率约为15.38%。结论灰色马尔柯夫预测模型较单独使用GM(1,1)模型,精度有显著的提高,预测结果将为科学的管理医院提供可靠的依据。?

Objective Combining the direct method of GM（1,1） and the principle of Markov forecast,we propose a Grey-Markov forecast model,and forecast the number of discharge in order to providing a statistics basis for management decision of hospital. Methods Based on the number of discharge in one hospital of Fujian from 2005 〈br〉 to 2015,using MATLAB,we applied the direct method to build GM （1,1） of discharge,then combined the principle of Markov forecast to partition three state interval of origin data and determined the future state by transition probability matrix,and obtained the finial prediction values. Results The relative error of prediction values from 2013 to 2015 were 0.05%,0.35% and 0.14% respectively,the prediction value of 2016 was 108335 person-time and the growth rate was 15.38%. Conclusions Compared to the direct method of GM （1,1）,Grey-Markov forecast model had better numerical performance in forecast accuracy. Based on forecast results,the reliable evidence would be provided for scientific managing hospital.