Forecast of Commercial Building Load Based on RF-CBR

The load forecast of commercial buildings is very important for power dispatching and building energy management, which can reduce building energy consumption. This paper proposes that importance scores of features can be obtained based on the Random Forest (RF) algorithm in the training process, which can be used as an important basis for the selection of feature variables. Case Based Reasoning (CBR) can make reasoning judgment on emerging cases based on existing cases and get the solutions of new cases. In this paper, the features for predicting the energy load of the target building are selected based on RF for case-based reasoning modeling, and the score of features importance for variables is used as the weight of the similarity calculation in the case-based reasoning process. The experimental results show that the modeling target can be achieved.