多因素数据重心预测法对比研究——中国钢材消费量预测实证分析

Comparison of multi-factor data barycenter forecasting method——Based on empirical analysis of forecast of steel material consumption in China

将多因素数据重心预测法与传统的最小二乘法应用于我国钢材消费量的预测,并从建模到预测结果的统计检验指标等方面进行对比研究。结果表明,多因素数据重心预测法在预测结果的稳健性方面具有较大的优越性,同时多因素数据重心法在处理具有异常数据的样本数据方面具有明显的平滑系统误差的作用。另外,多因素数据重心法还具有更直观、易于理解与运用、对数据的要求比最小二乘法宽松等优点。

Multi-factor data barycenter forecasting method and Least square method are applied to forecast Chinese steel consumption,and comparison research is made on forecasting result,statistical index on these two methods.The comparison results show that the stability of the forecasting result of multi-factor data barycenter forecasting method is better than that of least square method.And multi-factor data barycenter forecasting method has good effect on smoothing the system error on forecasting result with the sample data on the irregular dada.Moreover,multi-factor data barycenter forecasting method is easier to study and apply.