摘要
建立Holt-Winter预测模型,将具有线性趋势、季节变动和随机波动的时间序列进行分解研究,并与指数平滑法相结合,分别对长期趋势、趋势的增量和季节波动做出估计。介绍该模型的3个平滑方程和1个预测公式,以及初始值的计算和最优加权系数值的确定。通过实证分析,将预测结果与常用模型进行比较,说明该模型的预测精度较高。
Establishment of the Holt-Winter forecast model could make decomposing study on time series which has linear trend, seasonal change and random wave, and separately estimate the long-term trend, increase of trend and seasonal wave with combining the exponential smoothing. The essay introduces three smoothing equations and one forecast equation, calculation of original value and determination of optimized weight. Through case study and comparing forecast results with that of common models, the essay points out that the Holt-Winter model has higher forecast accuracy.
出处
《铁道运输与经济》
北大核心
2007年第1期79-81,86,共4页
Railway Transport and Economy