Abstract
Time series analysis is a method built in a particular time sequence for prediction. One of the models in time series analysis used for prediction is the ARIMA model introduced by Box and Jenkins. As time goes by, the ARIMA model was developed by applying algorithms, one of which was the Kalman Filter algorithm. This study aims to estimate the parameters of the ARIMA model used as the Kalman Filter’s initial value to forecast rainfall using ARIMA and ARIMA Kalman Filter. Determination of the ARIMA model is done by dividing the data into training and testing. The results obtained from the three training data have the same model, namely ARIMA (0,0,0) × (0,1,1)12 models but with different parameter values than those used as initial values for the Kalman Filter. The results obtained using the ARIMA model with Kalman Filter significantly affect the initial data of 90% training data model parameters with an RMSE value of 155,13. Then predictions are made, the results obtained by ARIMA Kalman Filter can follow the actual data, but from June to October, the prediction results cannot approach the actual data. According to events in the field, June to October is the dry season, where rainfall is deficient
A Kalman Filter Time Series Analysis Method for InSAR
