新疆能源需求总量预测——基于TEI@I方法论

Total Energy Demand Forecasting in Xinjiang——Based on TEI@I Methodology

新疆未来的发展迫切需要能源工业的支撑。掌握能源需求的预测方法,通过正确的方法提高能源需求预测的准确性,对于新疆能源资源的规划与配置、新疆能源发展战略的制定以及促进新疆经济社会发展意义重大。本文以TEI@I方法论为基础,提出新疆能源需求预测的研究框架,通过建立ARIMA模型、灰色GM(1,1)模型以及BP神经网络对新疆"十二五"期间能源需求总量进行预测;并以各种方法误差率的相对大小作为方法精确度的衡量标准,得到各种预测方法的组合预测权重,对新疆未来能源需求进行集成预测。结果表明,集成预测方法所得到的结果更加稳定和准确。

Based on TEI@ I methodology put forward by Wang Shouyang, this paper proposes an energy demand forecasting study framework in Xinjiang. We set up ARIMA model, Gray Model （1,1）, BP Neural Network to forecast the energy demand during the 12th Fifth - year Plan in Xinjiang respectively. The ratio of residual is used to integrate these models and we also foreeast the yearly energy demand for the next ten years. Forecasting resuhs show that the integrated result can make the forecast more stable and credible.