Research on heat transfer load prediction model based on XGBoost integrating temporal feature enhancement and particle swarm optimization

The accurate prediction of heat transfer load is crucial for the intelligent advancement of the heating industry. However, existing approaches often fall short in fully leveraging temporal characteristics and suffer from inefficient hyperparameter tuning. To overcome these limitations, this paper introduces a hybrid model that integrates a Temporal Feature Enhancement (TFE) algorithm with Extreme Gradient Boosting (XGBoost). The TFE algorithm reconstructs input features to better capture temporal dynamics, while Particle Swarm Optimization (PSO) is employed for efficient hyperparameter search. Following a structured pipeline of data collection, preprocessing, feature construction, modeling, and optimization, the model was validated using real-world data from a residential heat exchange station in Dalian. Experimental results demonstrate the model's exceptional performance, achieving a determination coefficient of 0.9998 alongside significant reductions in Root Mean Square Error (RMSE) and Mean Absolute Error (MAE). This confirms the model's superior capability in extracting profound patterns from time-series data.