Peramalan Konsumsi Gas Indonesia Menggunakan Algoritma Fuzzy Time Series Stevenson Porter

Abstract

There are some techniques of soft computing that can be used to forecast the data, they are fuzzy time series, neural network, and genetic algorithm. The methods can solve the data forecasting in the complex model that related to non linear time series model. In this research, the forecasting method that used is the Algorithm of Fuzzy Time Series Using Percentage Change As Universe Discourse that proposed by Stevenson and Porter who will compared with one of the classical forcasting method. The data that used is data of Indonesia gas consumption years 1990-2013, where the data contain trend pattern. For that, one of the good classical forecasting method that can be used is the method of Double Exponential Smoothing Holt. The results, forcasting by using the Algorithm of Fuzzy Time Series Stevenson Porter is better compared with using the method of Double Exponential Smoothing Holt, because it has smaller MAPE and MSE value, each of them in a row are 6,56 and 4,93. Where obtained the forecasting result for year 2014 is 38,39