PERKEMBANGAN MODEL PREDIKSI INDEKS K GEOMAGNET

Abstract

Prediction model of daily K geomagnetic index was reconstructed based on K index pattern of 3-hourly data from Biak geomagnetic station during 1992. K index was observed according time sequence so that it is very complex, fluctuated, and without any pattern, due to stochastic properties. With such data, K index prediction model was analyzed based on the smoothing result by using a moving average of four data in time series analyzsis method through ARIMA (Auto Regressive Integrated Moving Average) model. The obtained prediction model of K index, from moving average of four data from the first 3 hours to the next eighth 3 hour, generally follow the ARIMA (2.0.0), ARIMA (2.0.1) and ARIMA (2.0.2) with very small error, less than 1((σ error than more 1) and the efficiency of 68.27%, 94.10% and 95.16%, respectively, during year 1996.