Research on post-evaluation model and method of electric power demand forecasting

According to the complexity and stages of electric power demand forecasting, this paper established the hierarchical post-evaluation model based on the forecasting precisions of indexes in each layer, which could balance the different guidance to electric power development provided by the electric power demand forecasting and the adjustment forecasting. The model could also solve the problem which comprehensive evaluation could not acquire evaluation results in each layer. Besides, as there were only few cases of post-evaluation for 5-year demand forecasting in electric power industry, and many factors affecting the demand forecasting, this paper introduced a concept of weighted mutual information content. Thus an approach to allocate impersonal weights of factors affecting electric power demand forecasting was obtained. Case study results showed that the established model and proposed approach were effective and practicable.