Search for a command to run...
Energy efficiency, demand management, sustainable planning of an energy infrastructure, and successful forecasting of the electricity load in smart cities are very important. The current paper offers a new deep learning approach, ST-DEForecastNet, to this challenge of spatio-temporal forecasting in the context of urban building environments. The architecture consists of a temporal encoder relying on Temporal Convolutional Network (TCN) and Gated Recurrent Units (GRU), and a spatial encoder that uses Graph Attention Networks (GAT) to capture the intra-building energy correlations. Multi-head self-attention fusion layer achieves cross-dimensional dependencies, and multi-horizon predictions are provided by a residual forecasting head. The chosen publicly available dataset, Individual Household Electric Power Consumption, is necessary to appraise the suggested model. ST-DEForecastNet achieves better performance than other current models in terms of many metrics. As the accuracy level is recorded to be 95.1 maximum, it also has MAE, RMSE, and maximum R 2 of 0.124, 0.162 and 0.96, respectively. Also, the model has an AUC score of 0.97, which shows its high classification ability in predicting the consumption patterns. Whereas training and inference time are slightly up compared to lightweight models, it is well worth it according to calculations made in increasing forecasting accuracy and comprehension. The model has imbedded attention schemes that give visibility to critical time slot durations and sub-metered energy zones at which appliances have an effect on the energy consumption, making it appropriate in policy and operations of the smart energy system. ST-DEForecastNet has a modular and understandable structure, a scalable approach to forecasting energy demand in a complicated urban environment.