Sunspot Time Series Forecasting using Deep Learning

Abstract

In order to forecast solar cycle 25, sunspot numbers(SSN) from 1700 ∼ 2018 was used as a time series to predict the next eleven years. deep long short-term memory(LSTM) was exploited to do the forecast, first the dataset was split into training set(80%) and (20%) for the test set, the achieved accuracy led us to forecast the next eleven years. The result shows that the cycle will be from 2019 ∼ 2029 with peak at 2024.