基于季节性神经网络的医院门诊量曲线拟合与预测

医院月门诊量是一个具有复杂的非线性组合特征的季节性时间序列。传统的时间序列趋势分析是通过季节调整建立预测模型,效果不理想。文章提出一种利用季节性神经网络预测模型对医院门诊量进行非线性曲线拟合分析并预测。论述了该模型的设计思想和实现算法。通过仿真实验表明,该模型的非线性曲线拟合精度和预测精度明显高于ARIMA季节乘积模型,可较好地反映系统的动态性和门诊量的季节时序关联性,为季节性时间序列预测提供了一种新的途径。

Curve Fitting and Forecasting of Outpatient Amount Based on Seasonal Neural Network

The hospital outpatient amount is a seasonal time series with the character of complex nonlinear combination. The forecasting results based traditional forecast model with seasonal adjustment are not satisfied. This paper puts forward a seasonal neural network model to curve fitting analysis for nonlinearity and predict for the seasonal time series of outpatient amount. The paper discusses the designing ideas and realizing algorithms of the model. The simulation samples show that the precision of the curre fitting nonlinearity and prediction of this model is better than that of ARIMA seasonal product model. The model can reflect the dynamic characteristic of system and the seasonal time series connected characteristic of outpatient amount. Thus, a new way is obtained to predict the seasonal time series.