Comparison Of ARIMA And Exponential Smoothing Holt-Winters Methods For Forecasting CPI In The Tegal City, Central Java

Abstract

The Consumer Price Index (CPI) is an essential economic index that shows the level of prices for goods and services consumed by the public in a certain period in a specific region so that forecasting the ICP is needed to find out the pattern of economic movement in the area. The purpose of this study is to determine the forecasting rate for CPI from July 2021 to June 2022 by comparing two forecasting methods, i.e., ARIMA and Exponential Smoothing Holt-Winters. The data used in this study is Tegal City CPI data for January 2014 - June 2021, with the year 2018 as the base year equals 100 with a time series of 90 observations. The backcasting technique was implemented to the CPI figures of January 2014 – December 2019 (Base Year 2012=100) to adjust the new Base Year following 2018 on Classification of Individual Consumption According to Purpose (COICOP). The results from the two methods show that the Exponential Holt-Winters method has a minor Mean Absolute Percentage Error (MAPE) value, which is 0.281 compared to the MAPE of ARIMA value of 0.311. Hence, the Exponential Holt-Winters Additive method is chosen as the best CPI forecasting model for Tegal City