PERAMALAN DENGAN METODE SARIMA PADA DATA INFLASI DAN IDENTIFIKASI TIPE OUTLIER (Studi Kasus: Data Inflasi Indonesia Tahun 2008-2014)

Abstract

Inflation is defined as rising prices of goods in general and continuously. The effect of inflation on the economy can cause the currency to decline, resulting in the country's economic power becoming weak. Time series data is data arranged in order of time or data collected over time. Changes in the inflation rate tend to make inflation data unstable and affect the forecasting process in the time series data. The method used in this study is the seasonal autoregressive integrated moving (SARIMA) method to predict the time series in one or two periods ahead. This study also used outlier identifiers on models that still have outlier tendencies in residuals. The forecasting results of the SARIMA method become inaccurate when residual data contains outliers. The presence of outlier data in residual data results in residuals is not a normal distribution. The method used obtained the best model results, namely the SARIMA model (0,1,1) (0,1,1)12 with inflation forecast value for January to May 2015 is in the range of 5-6 %. On SARIMA models (0,1,1) (1,1,1)12 and SARIMA models (1,1,0) (2,1,0)12 outliers are detected in residual are Additive Outlier (AO) and Temporary Change (TC) type.