基于BP神经网络的公共建筑用电能耗预测研究

数据挖掘技术与建筑工程之间的知识跨度较大，将两者完美融合起来存在一定难度，实际工程中也缺乏相关案例，因此在建筑工程中应用数据挖掘技术挖掘相关信息，为大型公共建筑用电能耗预测提供参考依据，是建筑领域一种新的发展趋势。基于对公共建筑用电能耗特性的分析，可利用Python构建BP神经网络建筑能耗预测模型，再将某公共建筑作为研究对象，确定影响建筑用电能耗的关键因素，并将其作为网络的输入参数进行学习训练得出预测值。研究结果表明：预测模型在不同参数下，当隐含层个数为8时，误差平方和最小，为0.000 139 6，此时BP神经网络能够较精确地预测公共建筑用电能耗值。

Research on Prediction of Electricity Consumption of Public Buildings Based on BP Neural Network

The knowledge span between data mining technology and construction engineering is large. It is difficult to integrate the two perfectly. There is no relevant case in actual engineering. Therefore， using data mining technology to mine relevant information in construction engineering is used for large public buildings. The reference for electric energy consumption prediction is a new development trend in the construction field. Based on the research and analysis of the energy consumption characteristics of public buildings， we use Python to construct the BP neural network building energy consumption prediction model， and then a public building is taken as the research object to determine the key factors affecting the building energy consumption， and learning training is performed as an input parameter to the network. The results show that under the different parameters， when the number of hidden layers is 8， the sum of squared errors is the smallest， which is 0.000 139 6. At this time， the BP neural network can accurately predict the energy consumption value of public buildings.