利用2006年1月至2016年10月煤炭价格周数据,基于多重分形消除趋势法建立电煤价格波动特性分析模型,分析不同周期划分情况下电煤价格的波动规律。研究结果表明:我国电煤价格自身波动特性明显,局部波动特性同长周期电煤价格波动特性相似。同时,基于敏感性分析法对多重分形波动因子对电煤价格预测的修正效果进行了检验,证明因子可有效判断价格的剧烈波动趋势,预测误差比传统方法明显降低,可为政府制定相关煤电行业政策提供参考。 Based on the coal price weekly data from January 2006 to October 2016, an analysis model of coal price fluctuation was established based on the multifractal elimination trend method, and the fluctuation of coal price under different periods was analyzed. The results show that: the fluctuation characteristics of coal price in our country are obvious; the characteristics of local fluctuation are similar to those of long-period coal. At the same time, based on the sensitivity analysis, the multifractal fluctuation factor is tested to the correction effect of coal price forecasting. It is proved that the factor can effectively judge the trend of price volatility. The prediction error is obviously lower than the traditional method, Industry policies provide a reference.