Energy demand prediction method based on combined model and application thereof

The invention provides an energy demand prediction method based on a combined model and an application thereof. The prediction method comprises the following steps: predicting an energy demand by adopting an unbiased grey prediction model; a partial least squares regression model is adopted to predict the energy demand; determining the weight of each prediction method by adopting an entropy weight method; and adopting the combined prediction model to carry out weighted combination on each prediction model result to obtain a final prediction result. According to the method, effective information in a single model can be extracted, and errors caused by defects of different single models are reduced, so that the prediction precision is improved, and reasonable and effective prediction of energy demands is realized. 本发明提供了一种基于组合模型的能源需求预测方法及其应用，所述预测方法包括：采用无偏灰色预测模型预测能源需求；采用偏最小二乘回归模型预测能源需求；采用熵权法确定各预测方法权重；采用组合预测模型对各预测模型结果进行加权合并得到最终预测结果。本发明方法可以提取单一模型中的有效信息，降低不同单一模型的缺点带来的误差，从而提高预测精度，实现能源需求的合理有效预测。