Solar Cycle Prediction Using a Temporal Convolutional Network Deep-learning Model with a One-step Pattern

Abstract The human living environment is influenced by intense solar activity that exhibits periodicity and regularity. Although many deep-learning models are currently used for solar cycle prediction, most of them are based on a multistep pattern. In this paper a solar cycle prediction method based on a one-step pattern is proposed with the temporal convolutional network neural network model, in which historical data are input and only one value is predicted at a time. Through an autoregressive strategy, this predicted value is added to the input sequence to generate the next output. This process is iterated until the prediction of multiple future data is achieved. The experiments were performed on the 13 months–smoothed monthly total sunspot number data sourced from WDC-SILSO. The results showed that a one-step pattern fits solar cycles 20 to 25 well. The average fitting errors are mean absolute error = 1.74 and RMSE = 2.34. Finally, the intensity of solar cycle 25 was predicted with a one-step pattern. The peak was predicted to occur in 2024 October with a magnitude of 135.3 and end in 2030 November. By comparing the prediction results with other methods, our method is more reasonable and better than most methods. The codes are available on GitHub ( https://github.com/zhaocui1207/solar-cycle-prediction-by-tcn ) and Zenodo (doi: 10.5281/zenodo.14211884 ).