Assessment and Prediction of Energy Usage Based on ECV and ARIMA-DBN Methods

In the background of the current growing energy demand, traditional energy usage assessment and prediction systems, which are typically based on government-released energy consumption indicators, are difficult to achieve timely and accurate monitoring and prediction of energy usage. Therefore, an energy usage assessment and prediction method based on the entropy weighting method, criteria importance through inter-criteria correlation method with coefficient of variation method (ECV) combination methods and the auto-regressive integrated moving average with the deep belief network (ARIMA-DBN) combination methods is proposed. Firstly, an energy usage evaluation method based on the ECV combination approach is proposed, which achieves a comprehensive assessment of energy usage by constructing an indicator system. Then, an energy consumption prediction method based on the ARIMA-DBN combination approach is proposed to forecast the dynamics and development trends of energy usage. Finally, the effectiveness of the proposed methods is validated using energy consumption data from energy-consuming enterprises in Zhejiang province of China. The simulation results demonstrate that the proposed methods can effectively evaluate energy usage and improve the accuracy of energy consumption prediction.