基于经济先行指标的省级电力市场需求分析  

Analysis of Provincial Power Market Demand Based on Economic Leading Indicators

作　　者：崔巍[1,2] 刘洋[1] 张米尔[2] 李亚军[3] 杨海峰[3] 

机构地区：[1]大连海事大学管理学院,辽宁大连116026 [2]大连理工大学管理与经济学部,辽宁大连116024 [3]重庆市电力公司计划部,重庆400014

出　　处：《水电能源科学》2012年第5期191-195,共5页Water Resources and Power

基　　金：国家自然科学基金资助项目(71072124);大连市2009年科技计划基金资助项目(2009A13GX057);大连海事大学2010年研究生教育教学改革基金资助项目;2012年度中央高校基本科研业务费专项资金资助项目

摘　　要：针对经济变量之间存在严重的多重共线性,且经济周期相对于电力市场需求的变化有一定的滞后性问题,以重庆市为例,采用数据分析法筛选出经济先行指标,在此基础上建立向量自回归模型,并进行脉冲响应分析获得各先行指标的先行期数,再以逐步回归法对各先行指标进行分析,去除多重共线性,筛选出显著影响大的指标,得到省级电力市场需求预测模型,该模型以中国生铁产量、重庆市钢材产量、重庆市城乡居民储蓄存款、重庆市企业存款、美国货币供应量M2共5项指标作为解释变量。经统计学和计量经济学准则检验,模型具有科学性和准确性。On account of the serious multiple-collinearity in economic variables and the hysteresis of economic cycle as to the power market demand,taking the data of Chongqing for an example,vector auto regression model is established on the basis of the economic leading indicators.And it conducts impulse response analysis to obtain the leading periods.Then stepwise regression is used to analyze the leading indicators and it removes multiple-collinearity.At the same time,the indexes with significant influence are selected.So,the provincial power market demand forecasting model is established,which takes China pig iron production,Chongqing steel production,Chongqing urban and rural residents＇ deposits,Chongqing enterprise deposits and the United States M2 as the explain variables.The scientificalness and accuracy of the model is verified by test of statistics and econometrics.

关 键 词：先行指标 先行期数 电力市场需求 逐步回归 预测模型 

分 类 号：TM715[电气工程—电力系统及自动化]