GMDH网络在短期负荷预测中的应用

文章提出基于自组织方法的GMDH(Group Method of Data Handling)型神经网络并将它应用于短期负荷预测.与一般的前馈神经网络不同,GMDH网络的结构确定于训练过程之中,因而可大大提高神经网络性能.它能充分、合理地利用数据,自动进行变量组合,筛选及判断从而得到合适的模型,特别适用于数据预测.将这种用自组织方法所构成的GMDH型神经网络应用于广西某地区电力局的短期负荷预测,采用Matlab6.5进行仿真实验,证明其在短期负荷预测方面有很好的应用前景.

The Applications of GMDH Neutral Network to Short-term Power Load Forecasting

This article puts forward GMDH (Group Method of Data Handling) neutral network based on serf-organizing method and its applications to short term power load forecasting. Differing from free-forward neutral network, GMDH neutral network can greatly enhance its performance as the structure of GMDH neutral network is set during the drilling process. GMDH network is especially suitable for data forecasting as it can reasonably make full use of data, combine variables auto-matically, screen and judge and identify the right model. This article is to testify its rosy allocation vistas in short term power load forecasting through emulation experiment via Matlab6.5 and application of GMDH in short term power load forecasting in some electric power bureau of Guangxi.