Prediction of Energy Consumption Time Series Using Neural Networks Combined with Exogenous Series

Artificial Neural Networks (ANNs) are widely used in various practical problems about time series. In this paper, a methodology based on exogenous series is used in combination with a Back Propagation (BP) neural network to predict time series. Exogenous series is chosen by correlation theory with endogenous series. In this way, the prediction output is obtained by not only the historical data but also the information external to historical data. Communication base station energy consumption is one important part of the total social energy consumption. So its energy consumption time series (ECTS) is used as the research data. We compare the prediction performance with the normal time delay neural network (TDNN), and the experiments show that the new method has a more precise and stable performance.