AN EXPONENTIAL AUTOREGRESSIVE MODEL FOR THE FORECASTING OF ANNUAL SUNSPOTS NUMBER
Vol. 7(3) August 2019, No. 3, pp. 17-23
NABIL AZOUAGH AND SAID EL MELHAOUI
We consider the problem of modelling the time series of annual sunspots number because of its effect on earth temperature. Indeed, sunspots are colder than the Sun’s surface and therefore decrease slightly the intensity of solar radiation. This small variation represents a significant difference in energy received for a planet like Earth. In this work, we used the Exponential Autoregressive model (EXPAR) for the forecasting of the annual sunspots number. We justified the use of the EXPAR model by applying a non-linearity test for the detection of the exponential component. The prediction errors resulting from the EXPAR model were less than those of other linear and non-linear models considered in the literature for modelling the sunspot time series.