A nonlinear approach to modelling the residential electricity consumption in Ethiopia
A nonlinear approach to modelling the residential electricity consumption in Ethiopia
复制标题
埃塞俄比亚住宅用电量建模的非线性方法
DOI：
10.1016/j.eneco.2009.08.008
复制
发表时间：
2010
期刊：
影响因子：
12.8
通讯作者：
Emmanuel Gabreyohannes
中科院分区：
文献类型：
--
作者：
Emmanuel Gabreyohannes
本文尝试用自激励门限自回归(SETAR)模型和平稳过渡回归(STR)模型对埃塞俄比亚居民生活用电量进行建模、分析和预测。为了进行比较，该应用程序也被扩展到标准线性模型。在两个模型的实证演示过程中，都发现了显著的非线性效应，并拒绝了线性。在样本点外和区间(密度)预测方面，SETAR模型相对优于线性自回归模型。我们的STR模型的结果表明，拟合的STR模型的残差方差仅为线性ARX模型的65.7%。因此，我们可以得出结论，包括非线性部分，这基本上解释了极端价格事件的到来，导致了模型对埃塞俄比亚电力消费的解释能力的提高。
In this paper an attempt is made to model, analyze and forecast the residential electricity consumption in Ethiopia using the self-exciting threshold autoregressive (SETAR) model and the smooth transition regression (STR) model. For comparison purposes, the application was also extended to standard linear models. During the empirical presentation of both models, significant nonlinear effects were found and linearity was rejected. The SETAR model was found out to be relatively better than the linear autoregressive model in out-of-sample point and interval (density) forecasts. Results from our STR model showed that the residual variance of the fitted STR model was only about 65.7% of that of the linear ARX model. Thus, we can conclude that the inclusion of the nonlinear part, which basically accounts for the arrival of extreme price events, leads to improvements in the explanatory abilities of the model for electricity consumption in Ethiopia.