Predicting electricity demand in China rural considering new-type industries’ multidimensional features

With the implementation of China’s rural revitalization strategy, an increasing number of new-type industries have invaded the rural areas, thereby enhancing uncertainty of rural electricity consumption. In this context, traditional methods for electricity demand forecasting are no longer able to meet new features of the evolving rural electricity consumption. For this sake, this study proposes a new prediction model integrating long short-term memory (LSTM) network and principal component analysis (PCA), considering multidimensional features and multimodal data of new-type industries’ electricity demand. Through comparing with other typical electricity prediction models, the novel model is shown to have significant advantages in dealing with the high volatility characteristics of data with prediction accuracy. This study provides effective tools and decision support for sustainable improvement of rural power supply and service levels.