PERBANDINGAN METODE DESEASONALIZED ARMA DAN METODE SARIMA DALAM PERAMALAN RUNTUN WAKTU MUSIMAN : Studi Kasus pada Rata-rata Harga Beras di Tingkat Perdagangan Besar atau Grosir Indonesia

• Main Author: Yasmine Nabillah, - (Author)
• Format: Book
• Published: 2019-08-21.
• Subjects: Thesis; NonPeerReviewed
• Online Access: Link Metadata
• Notes: http://repository.upi.edu/39416/9/S_S_MAT_1503713_Title.pdf; http://repository.upi.edu/39416/8/S_S_MAT_1503713_Title.pdf; http://repository.upi.edu/39416/2/S_S_MAT_1503713_Chapter1.pdf; http://repository.upi.edu/39416/1/S_S_MAT_1503713_Chapter2.pdf; http://repository.upi.edu/39416/7/S_S_MAT_1503713_Chapter3.pdf; http://repository.upi.edu/39416/3/S_S_MAT_1503713_Chapter4.pdf; http://repository.upi.edu/39416/6/S_S_MAT_1503713_Chapter5.pdf; http://repository.upi.edu/39416/4/S_S_MAT_1503713_Appendix.pdf
• Keywords: L Education (General); QA Mathematics

Ramalan dapat diperoleh dengan bermacam‐macam cara yang dikenal dengan metode peramalan. Runtun waktu musiman merupakan runtun waktu yang mengandung variasi musiman. Beberapa metode peramalan yang dapat digunakan untuk meramal data runtun waktu musiman yaitu metode Deseasonalized ARMA dan metode Seasonal Autoregressive Integrated Moving Average (SARIMA). Metode Deseasonalized ARMA merupakan metode dimana komponen musiman dihilangkan dari data runtun waktu yang selanjutnya dimodelkan dengan menggunakan Autoregressive Moving Average (ARMA) dan untuk peramalan maka komponen musimannya akan dikembalikan. Metode SARIMA merupakan pengembangan dari metode ARIMA. Kedua metode diaplikasikan pada data rata-rata harga beras pedagangan besar Indonesia dari Januari 2011 sampai Maret 2019, sehingga diperoleh ramalan harga beras untuk bulan-bulan berikutnya. Penelitian ini membandingkan antara model Deseasonalized ARMA yaitu ARIMA (1,1,0) dan model SARIMA yaitu ARIMA (1,1,0)(0,1,0)12. Hasil penelitian menunjukkan bahwa model peramalan rata-rata harga beras lebih baik menggunakan metode Deseasonalized ARMA dengan MSE sebesar 107.604,42 dibandingkan dengan SARIMA dengan MSE sebesar 284.085,13. Kata kunci: Peramalan, Runtun Waktu Musiman, Deseasonalized ARMA, SARIMA. The forecast could be obtained with some process known as forecasting methods. Seasonal time series is a time series that contain seasonal variation. Some methods that could be used to forecast seasonal time series are Deseasonalized ARMA method and Seasonal Autoregressive Integrated Moving Average (SARIMA) method. Deseasonalized ARMA is a method where the seasonal variation are removed from the actual time series and modeled using Autoregressive Moving Average (ARMA) and for forecasting the seasonal variation is returned. SARIMA method is an extension to ARIMA. These two methods are applied to average price of rice at Indonesian large trade level from January 2011 to March 2019, to forecast the average price of rice for the following months. This research is comparing Deseasonalized ARMA model that is ARIMA(1,1,0) and SARIMA model that is ARIMA(1,1,0)(0,1,0)12. This research showed that the average price of rice model is better modeled by Deseasonalized ARMA method with the measure of its MSE 107.604,42 which is smaller than MSE of SARIMA method 284.085,13. Key words: Deseasonalized ARMA, SARIMA, Rice Price, Seasonal Time Series