A prediction Algorithm Of Medium And Long Term Electricity Consumption Trend Considering The Characteristics Of Industry Electricity Consumption

Abstract

Different industries have their own characteristics of power consumption, and the development law of power consumption is obvious. When the existing forecasting algorithms assign weights to the combination model, the multi entropy distribution tends to be close to 1, which leads to large error in the forecasting results. Aiming at this problem, a medium and long-term electricity consumption trend forecasting algorithm considering the characteristics of industrial electricity consumption is proposed. According to the similarity of industry power consumption time series, the industry classification system is divided again. From the perspective of industry dimension and time dimension, the multi-level coordination model of medium and long-term electricity consumption is constructed, and the single weight of multi-level coordination combination is calculated by using the improved entropy weight method, so that the coordination model can reasonably allocate the single weight of combination model. The features of each component of the coordination model are extracted, and combined with the classification results of power consumption industry, the medium and long-term power consumption trend prediction algorithm is designed. The experimental results show that the algorithm designed in this paper has smaller prediction relative error, and the prediction results are in good agreement with the actual power consumption, which is helpful to improve the refinement level of medium and long-term power consumption trend prediction.