Sunspot interval prediction based on fuzzy information granulation and extreme learning machine
- PDF / 989,356 Bytes
- 11 Pages / 595.276 x 790.866 pts Page_size
- 8 Downloads / 157 Views
J. Astrophys. Astr. (2020)41:29 https://doi.org/10.1007/s12036-020-09649-4
Sadhana(0123456789().,-volV)FT3](012345 6789().,-volV)
Sunspot interval prediction based on fuzzy information granulation and extreme learning machine PENG LINGLING School of Artificial Intelligence, Chongqing University of Arts and Sciences, Chongqing, China. E-mail: [email protected] MS received 15 April 2020; accepted 17 August 2020 Abstract. Sunspot prediction is an important task for space weather and solar physics. Traditional point forecast may not be sufficiently satisfactory and reliable. To quantify the uncertainty of point prediction, a hybrid interval prediction model has been proposed for sunspot forecasting. Three major steps are taken: (1) the complementary ensemble empirical mode decomposition (CEEMD), to decompose the sunspot sequence into a series of modal components, (2) the fuzzy information granulation (FIG), to extract the minimum, average and maximum value of each window, and (3) the extreme learning machine (ELM), to conduct point prediction and interval prediction, superimposing the prediction values of all components as the final forecast results. The empirical study focus on the 13-month smoothed monthly sunspot number recorded by Solar Influences Data Analysis Center (SIDC) and show that the mixed model with the filtered CEEMD is more effective than the unfiltered one. It also enables us to track changes of the sunspot number with fast calculating speed and high accuracy both in point prediction and interval prediction. Keywords. Interval prediction—point prediction—complementary ensemble empirical mode decomposition—fuzzy information granulation—extreme learning machine—sunspot.
1. Introduction The temporal and spatial evolution of the magnetic activity on the surface of the solar atmosphere is closely related to the ecological environment, and its changes directly affect space navigation, radio communications, atmospheric movements, marine activities and Earth’s climate (Hiremath 2006; Hathaway 2015; Pala & Atici 2019). Sunspot is not just an important indicator of solar activity, it is also the most easily observed phenomenon in solar activity (Pucha et al. 2016; Gonc¸alves et al. 2019; Tan 2018). So how to accurately predict the sunspot number has always been an important subject. At present, domestic and foreign scholars have conducted extensive research on the prediction for the 13-month smoothed monthly sunspot number (denoted as SSN) (Miao et al. 2015; Yin & Han 2018; Javaraiah 2019). Among them, the forecast based on the precursory theory (Miao et al. 2020; Pesnell & Schatten 2018) and generator model
(Choudhuri 2015; Dikpati & Gilman 2008) has achieved a certain prediction effect. It is another important research field to analyze and study the historical observation data of sunspot by using a mature statistical model or artificial intelligence algorithms, such as exponential smoothing, entropy, wavelet analysis, time series method, artificial neural network and deep learning (Kakad et al. 201
Data Loading...
