摘要
年电力需求时间序列为既含有确定性的动态趋势又含有随机性波动的非平稳时间序列,对于平稳随机序列,自回归滑移平均(ARMA)是最成熟的统计学分析方法之一,而灰色系统理论(GM)则是一种动态趋势预测理论,将时间序列分析与灰色系统理论相结合用于年电力需求预测是一种非常有益的探索。先采用灰色建模从数据中得到趋势项的数学模型,然后对剔除趋势项之后的数据进行时间序列分析,建立ARMA模型,将以上两个模型结合起来构成组合模型,用于预测年电力需求。应用实例证明,该方法具有容易实现,预测准确的优点,是一种有效的预测方法。
The annual electric consumption time series is a non-steady time series that includes both the definite dynamic tendency and the random fluctuation. To the steady random series, autoregression moving average （ARMA） is one of the maturest statistics analysis methods; moreover the gray system theory is a kind of dynamic tendency forecast theory. Unification of the time series analysis and the gray system theory to forecast the annual electric consumption is a kind of extremely beneficial exploration. First use the gray modeling or stepwise regression to obtain the mathematical model of tendency term from the data, then process the data from which tendency term has been eliminated with time series analysis, building ARMA model, then unify the above two models to a combination model to forecast the annual electric consumption, The application example proves that the method can be realized easily and predict accurately. It is an effective forecast method.
出处
《广东电力》
2007年第2期10-13,共4页
Guangdong Electric Power
关键词
时间序列
灰色模型
数学模型
年电力需求
预测
time series
gray model
mathematical model
annual electric consumption
forecasting
