Residential Energy Consumption Prediction Model Based on BP Neural NETwork

Abstract

The energy consumption system of the residence is complex, and its factors are intertwined and influenced by various industries and fields. Aiming at a series of problems in the current society about residential energy consumption, uneven distribution of energy and waste, this paper establishes a prediction model of residential energy consumption based on BP neural network. Based on the historical residential energy consumption data of the UCI database, modeling and simulation with the help of MATLAB software, BP neural network model is trained until the best accuracy is achieved, and the number and learning rate of the hidden layer neurons with the best fitting degree are obtained, and the test samples are substituted into the training. After comparing the predicted value of the operation result with the actual energy consumption, it is found that the prediction error of the residential energy consumption prediction model is small, and it has practical value for the research of residential energy consumption prediction.