Prediction Research on the Energy Consumption of Public Building Based on MLR-BP Neural Network

Abstract

Accompanying the constant development of data mining technology and neural network, the application of computer technology in energy consumption prediction of large public buildings is the trend of green buildings and intelligent life in the field of engineering. In this paper, a new energy consumption prediction algorithm based on MLR-BP is proposed, which solves the issue of current energy consumption algorithm: the massive amount of calculation and the low prediction accuracy. Firstly, factors that affect building energy consumption are categorized as external factors (meteorological factors) and internal factors. In addition, the multiple linear regression is used to model the external factors, and the influencing factors with high correlation are screened out. Then, the selected external factors and internal factors are combined into the training of BP neural network to obtain a new energy consumption prediction model. The simulation results show that in comparison with the BP neural network, the proposed prediction algorithm in this paper can predict the building energy consumption more accurately.