我国铁矿石进口价格预测的ECM-SVR混合模型

作为钢铁工业生产的重要原材料之一,铁矿石进口价格的剧烈波动给我国钢铁企业带来巨大的冲击.本文通过分析影响铁矿石价格波动的多种因素,包括供需关系、运费成本、国内外经济环境等,挖掘影响铁矿石进口价格的关键因素,综合考虑其线性和非线性均有的复杂时间序列特征,提出一种基于误差修正模型(error correction model,ECM)和支持向量回归(support vector regression,SVR)的铁矿石价格混合预测模型ECM-SVR.实证结果表明:与单一基准模型和传统混合模型相比,新模型具有较高的预测准确率,这对于钢铁企业控制原料成本和市场投资者合理规避价格风险具有重要指导作用.

A hybrid ECM-SVR model for iron ore price forecasting in China

As an important raw material for iron and steel industry, the fluctuation of iron ore price brings great impact and risk to iron and steel enterprises in China. For exploring the main factors affecting market price, a systematic analysis is conducted on iron ore supply, demand, freight costs and macroeconomic environment, taking into account the complexity of a time series which behaves both linear and nonlinear characteristics, a new hybrid ECM-SVR model based on error correction model and support vector regression is proposed for iron ore price forecasting. The empirical results demonstrate that the hybrid model has higher prediction accuracy compared with single models and traditional hybrid model. This study is instructive for the steel enterprises to control the cost of raw materials and for the market investors to avoid the price risk.