Electricity Consumption Forecasting Based on Improved Prophet Modeling

With the gradual increase in the importance of electricity consumption prediction, in order to guide the enterprise users to use electricity rationally, the management department realizes the monitoring and prediction of daily electricity consumption, for this reason, a new type of electricity consumption prediction model is proposed. Aiming at the problem that some electricity models lack the processing of special time nodes and are prone to overfitting, which leads to large deviations in the confidence interval of prediction, an improved Prophet model is proposed by integrating the Prophet model with the XGBoost model. The model increases the predictive value optimization block, which can effectively solve the problem of overfitting and achieve accurate prediction. The feasibility and effectiveness of the model in predicting the electricity consumption of users are examined in the context of a talent apartment in Tianjin to improve the accuracy of electricity consumption prediction.