Multi-Scale Power Load Forecasting Based on Adaptive GCN and Multi-Head Attention Mechanism. Si

Key Points

• The aim is to enhance the accuracy of power load forecasting by addressing non-stationarity and multi-scale features associated with renewable energy.
• Utilized an adaptive graph convolutional network (GCN) for data modeling.
• Implemented a multi-head attention mechanism to capture complex relationships in the data.
• Conducted various prediction analyses based on renewable energy integration.
• Demonstrated increased forecasting accuracy compared to traditional methods.
• Achieved improved handling of non-stationary data characteristics.
• Highlighted significant gains in multi-scale prediction performance.