The Implementation of a Filter Kalman Method Forecasting Rainfall Obtained Through Model ARIMA in Kota Jambi

In the last three years the climatic conditions in Jambi City have experienced erratic weather conditions. One way to predict rainfall is using the Kalman Filter approach. However, in this case, the Kalman Filter method is implemented on the forecasting results from ARIMA (Autoregressive Integrated Moving Average) because there has been rainfall measurement data from 2008 to 2017 at the BMKG Muaro Jambi Climatology Station which is also a function of time and the existing pattern will be described with using Time Series Analysis. Time series data is data that has a time series of more than one year on one object or data collected from time to time on one object. ARIMA model will be used to predict the next data. Kalman filter is a model part of state space that can be applied in forecasting models. The Kalman filter consists of a prediction stage and a correction stage. This method uses a recursive technique to integrate the latest observational data into the model to correct previous predictions and make further predictions. This study aims to determine the implementation of the kalman filter method in predicting rainfall obtained through the ARIMA model in Jambi City. The results of the 2018 Jambi City rainfall prediction research show that the best ARIMA model formed is the ARIMA model (1,0,1). In the Kalman Filter model, a MAPE value of 24.92% is obtained, which indicates that the Kalman Filter has a fairly good predictive ability.