组合预测在我国能源消耗总量中的应用.doc

组合预测在我国能源消耗总量中的应用 摘　要 组合预测模型是将各种不同类型的单项预测模型兼收并蓄，各取所长，集中了更多的经济信息与预测技巧，能减少预测的系统误差，显著改进预测效果．本文利用我国能源消耗总量的历史数据，建立灰色预测模型、移动平均预测模型和多元回归预测模型，根据单项预测得出的结果进行权重分配，建立组合预测模型，预测结果表明预测值和实绩分析结果有很好的一致性，可以作为研究我国能源消耗总量的有效工具．通过比较各种模型的预测误差和从定量的角度来提高预测精度，为预测我国未来的能源消耗总量提出了改进方向，并通过各种预测方法所提供收集有用的信息加以综合利用，提高了预测精度，为我国的能源消耗总量和利用提供了依据和理论基础． 关键词：组合预测 能源消耗 单项预测 预测精度 Application of Combination Forecast in Terms of Total Energy Consumption in China Zhang Kai Directed by Lecturer Jiang Shutao Abstract Combination forecast model is a different type of forecasting model of individual learning strengths, focus on a more economic information and forecasting skills, can reduce the forecast error, and significantly improve forecast. This project uses historical data of China’s total energy consumption, building grey forecasting models, the moving average forecasting model and multiple back-forecasting models, weight distribution according to the results of individual forecasts, the establishment of combined forecasting model, predictions showed that the predicted values and performance analysis results with good consistency, can be used as effective tools for study of China’s total energy consumption. Errors by comparing the various forecasting models and improve forecast accuracy from a quantitative perspective, projections suggest improvements for the future direction, and the useful information provided through the comprehensive utilization of various forecasting methods, significantly improve forecast accuracy, provided the basis for China’s energy consumption and the rational use and theoretical foundation. KEYWORDS: Combination forecast Energy consumption Items forecast Forecast accuracy 目 录 摘　要 I 英文摘要 II 前　言 1 1 单一预测在我国能源消耗总量中的应用 2 1.1灰色预测 2 1.1.1数据的选择 2 1.1.2构造累加生成列 2 1.1.3构造矩阵B和数据向量 2 1.1.4得出预测模型 3 1.1.5残差检验 4 1.2移动平均法 5 1.3多元回归预测模型 7 1.3.1建立模型 7 1.3.2估计参数 7 1.3.3进行检验 9 2 组合预测在我国能源消耗总量中的应用 9 2.1各单项预测的误差分析 9 2.2计算组合权重，建立组合预测模型 11 3 总 结 13