Improving Energy Demand Prediction Use Deep Learning Network

In recent years, the convergence of deep learning and energy demand forecasting, particularly for short-term power demand, has become a burgeoning field of research. As societies become increasingly reliant on electricity for various aspects of life, accurate forecasting has become critical for maintaining a stable and efficient electricity supply. In this study, we evaluate the performance of various deep learning algorithms to enhance the accuracy of short-term electricity demand forecasting. Additionally, we consider the impact of public concern, as measured by Google Trends, as an exogenous variable that affects electricity demand forecasts. The experimental results show that the hybrid model combining Convolutional Neural Network with Long Short-Term Memory (CNN-LSTM) architecture outperforms both the stand-alone LSTM and Sequence to Sequence model with Attention (Seq2Seq-Attention), providing a better prediction in both cases, whether including or excluding Google Trends data.