Metode Pemulusan (Smoothing) Eksponensial Ganda (Linier Satu Parameter dari Brown) dan Metode Box-Jenkins dalam Meramalkan Curah Hujan di Kota Medan

Abstract

DOUBLE EKSPONENTIAL SMOOTHING METHOD (LINIER ONE PARAMETER OF BROWN) AND BOX-JENKINS METHOD IN PREDICTING RAINFALL IN MEDAN CITY ABSTRACT In this research, method that is used in predicting rain pour are double exponential smoothing method (linier one parameter from brown) and Box-Jenkins ARIMA method. In double exponential smoothing is firstly finding a single exponential smoothing of the value that is got from single exponential smoothing, then, finding a value or adaptation of a single value and determining tendency estimation of a period of time to another that is b value and at last is doing prediction for next period. Meanwhile, box-jenkins ARIMA method is firstly doing differencing. It is for getting stationary data and next step is identifying model, estimating parameter of the model, doing verification parameter of the model, determining better model by looking at least error value of previous chosen model and finally making prediction for next period of time. Based on result of both prediction methods is got SSE and MSE brown value consecutively is 977.828.884 and 16.859,119. While SSE and MSE box-jenkins value consecutively is 665.432 and 15,475. So chosen method in this research is box jenkins ARIMA method because of less error value than brown method.

Keyword: Forecasting, Double Exponential Smoothing, Box-Jenkins, error