Comparison of Seasonal Time Series Forecasting Using Sarima and Holt Winter’s Exponential Smoothing (Case Study: West Sumatra Export Data)

Abstract

Export is the activity of selling goods or services from one country to another. This activity usually occurs in a specific region or country. Export data is a type of data that has a seasonal pattern. This study aims to compare SARIMA and Holt Winter’s methods in forecasting export data. In this study, the SARIMA model ((1,1,1) (0,1,1))12 and Holt Winter's simulation were obtained. The data used is the export data of West Sumatra from 2016 to 2022. The best model is the one with the smallest MAPE or MAD. The SARIMA model yielded a MAPE of 0,437% and MAD of 78,821. Meanwhile, the Holt Winter's method yielded a MAPE of 0,894% and MAD of 163,320 with α=0,2, β=0,5, γ=0,1. Therefore, the SARIMA outperformed the Holt Winter’s method due to its higher accuracy. It can be concluded that the SARIMA is suitable to use as the forecasting model in this case. In this study, forecast have been made for the next 24 periods, from January 2023 to December 2024.

Penerapan ARIMA Pada Data Curah Hujan di Stasiun Meteorologi Kelas II Minangkabau Padang Pariaman

“…Hasil penelitian dengan metode Box-Jenkins mengestimasi model berdasarkan nilai minimum AIC dan BIC atau salah satu dari keduanya [8]. Namun, sebelum itu estimasi parameter telah dilakukan di semua model yang diperoleh serta melakukan uji asumsi white noise dan asumsi distribusi normal terhadap residual [9].…”