Research on Identification Method of Electricity Consumption Behavior of Small and Medium-Sized Enterprises Based on Random Forest

With the increasing demand for data monitoring of micro, small and medium-sized enterprises and the rapid development of digital grid construction, power data resources have increased sharply, which contains rich information resources. However, in the data analysis of micro, small and medium-sized enterprises, due to industry differences, regional differences, difficulties in scale division and other reasons, there are usually problems such as the size and industry characteristics of micro, small and medium-sized enterprises are not accurate enough, and the data analysis error is large, which further affects the later decision-making and policy development. In order to solve the problem that the electricity consumption of micro, small and medium-sized enterprises is not accurate enough, this paper proposes a random forest-based identification method of electricity consumption of micro, small and medium-sized enterprises. Firstly, the power consumption data of small, medium and micro enterprises are collected and pre-processed to establish a decision tree model. Then, the random forest model is constructed, and the power consumption identification model is established by training the power consumption data of small, medium and micro enterprises. Finally, the simulation is carried out according to the actual operation data of small, medium and micro enterprises in a city in Northeast China. The results show that the method proposed in this paper can effectively identify the characteristics of small, medium and micro enterprises and ensure the accuracy of the solution, which is of great significance for improving the safe and stable operation and economy of small, medium and micro enterprises.