滑动数据重心预测方法及其应用研究

本文在研究多因素数据重心法的基础上,进一步提出滑动数据重心预测方法,该方法是对原始样本数据提出了一种新的数据处理方法,大大降低了由于历史数据组中的异常点对预测结果产生的破坏性。通过建立我国钢材消费量与国内生产总值(GDP)的计量动态模型对该方法与多因素数据重心预测法进行对比研究。同时利用时间序列自回归AR(p)对计量动态模型的初级预测结果进行差值校正,并将该方法应用于我国2015年、2020年的钢材消费量预测。对比研究表明该方法使得预测结果更加精确、稳健。

Research on Gliding Data Barycenter Forecasting Method and Its Application

A new Gliding data barycenter forecasting method is proposed on this paper based on Multi-Data Barycenter method,which is a new data processing method for original sample data and greatly reduce the destroy of singular point among the sample data group for the forecasting result.This new method is compared with Multi-factor Data Barycenter Method through establishing econometrics dynamic model between the relationship of China steel consumption and GDP on this paper.Autoregressive moving average AR(p) model is used to revised the primary forecasting results of the econometrics dynamic model,then China steel consumption in 2015 and 2020 year are forecasted based on the new method.Research results show that the new method made the forecasting result more accurate and robust.