Transformer based prediction method for solar power generation data

Abstract

In this paper, we propose a technique to increase the precision of solar power generation data prediction by using a time-series-based transformer deep learning model. By partially modifying the transformer model, which is widely used for language translation, we use it by changing the input and output of the model in the form of predicting future data. Finally, through comparison with other prediction models, it was confirmed that the proposed model showed very good prediction performance.