Conceptual Study of NNARX Method for Solar Radio Burst Prediction Model Development

Dependencies for electromagnetic spectrum-based technologies have increased with the introduction of digital broadcasting and high-speed internet. The solar weather which specifically related to the sun’s activities caused an impact toward the stability and the continuity of electromagnetic spectrum on earth with the strike of solar flare. With the use of low-frequencies radio spectrometers, the early detection of solar flare can be made with the appearance of solar radio bursts. With the prediction model developed for solar radio burst, the effect of it can be reduced with a precaution action. Artificial Neural Network (ANN) had been widely applied in various prediction model development. Since solar radio flux is a highly non-linear data, Neural Network Autoregressive Model with Exogenous Input (NNARX) is considered for the solar radio burst prediction model development. NNARX is a dynamic ANN which had been widely applied in solar radiation prediction model development. In this paper, previous study on the application of NNARX will be discussed to develop an architectural concept in order to prepare the NNARX for solar radio burst prediction. By the end of this study, framework for the prediction model development is established to prepare future work for solar radio burst prediction using NNARX.