Projection of energy consumption with artificial neural network at regional level: A case study of Chongqing municipality in China

Because the increasing energy demand have an important impact on economic development, the accurate projection of energy consumption is crucial to the energy policy decision. In this study, artificial neural network (ANN) model is used to estimate the energy consumption for Chongqing in China. The projection is implemented using a feed-forward neural network, trained by back-propagation algorithm. In order to investigate the socioeconomic influences on energy consumption, the ANN model is trained based on the gross domestic production and population along with the energy consumption data available from 1949 to 2008. The projection results from different models are compared to test the performance of model validation. It is found that the ANN model shows a better projection of energy consumption than the logistic and exponential model in terms of both the absolute mean relative error and the absolute fraction of variance. In short, the ANN model is a feasible method for projecting energy consumption so as to provide some implications relating to energy and economic policy.