Optimization of modeling method for building energy consumption prediction model

This paper optimizes the modeling method of building energy consumption prediction model. Optimized content includes data preprocessing and model framework selection. Data preprocessing consists of three steps. First, GESD algorithm is used to detect outliers in the data set. Then. K-shape algorithm clusters building energy consumption in the form of time series, which is rarely used in building energy consumption prediction models. Finally, the input variables are selected according to the partial mutual information algorithm.The use of information entropy can effectively remove the redundancy between various variables. The framework of the prediction model is the random forest algorithm. Random forest is an integrated learning algorithm, which can gather the advantages of each model. The single model used by most building energy consumption prediction models is not very effective. The final result shows that compared with other traditional prediction models, the building energy consumption prediction model proposed in this paper has higher prediction accuracy.