EMPIRICAL MODE DECOMPOSITION - SEASONAL AUTOREGRESSIVE INTEGRATED MOVING AVERAGE (EMD - SARIMA) (Studi Kasus: Peramalan Curah Hujan Kota Bandung)

• Main Author: Regina Kemala Pratiwi, - (Author)
• Format: Book
• Published: 2021-08-25.
• Subjects: Thesis; NonPeerReviewed
• Online Access: Link Metadata
• Notes: http://repository.upi.edu/64597/8/S_MAT_1700686_Title.pdf; http://repository.upi.edu/64597/2/S_MAT_1700686_Chapter1.pdf; http://repository.upi.edu/64597/3/S_MAT_1700686_Chapter2.pdf; http://repository.upi.edu/64597/4/S_MAT_1700686_Chapter3.pdf; http://repository.upi.edu/64597/5/S_MAT_1700686_Chapter4.pdf; http://repository.upi.edu/64597/6/S_MAT_1700686_Chapter5.pdf; http://repository.upi.edu/64597/7/S_MAT_1700686_Appendix.pdf
• Keywords: L Education (General); QA Mathematics

Penelitian ini bertujuan untuk meramalkan curah hujan di Kota Bandung berdasarkan data periode Januari 2015 sampai Desember 2019. Peramalan dilakukan dengan menggunakan metode Empirical Mode Decomposition - Seasonal Autoregressive Integrated Moving Average (EMD-SARIMA) yang merupakan gabungan metode dekomposisi EMD dengan metode runtun waktu musiman SARIMA. Metode ini digunakan untuk mengatasi permasalahan runtun waktu yang memiliki karakteristik nonlinear dan berpola musiman. Prosedur peramalan dengan metode ini diawali dengan uji nonlinearitas data. Hasil pengujian menunjukkan bahwa data nonlinear. Kemudian dilakukan dekomposisi dengan metode EMD. Hasil penelitian ini menunjukkan bahwa peramalan curah hujan di Kota Bandung untuk 12 bulan ke depan menunjukkan adanya fluktuasi di mana curah hujan tertinggi terjadi pada Bulan Agustus 2020 dan curah hujan terendah terjadi pada Bulan Januari 2020. Hasil peramalan masih jauh dengan data asli namun memberikan nilai MAPE sebesar 38,13307% yang artinya hasil peramalan tergolong cukup akurat. Kata Kunci: Nonlinear, Pola Musiman, EMD, SARIMA, Curah Hujan. This study aims to predict rainfall in Bandung City based on data from January 2015 to December 2019. Forecasting is carried out using the Empirical Mode Decomposition - Seasonal Autoregressive Integrated Moving Average (EMD-SARIMA) method, which combines the EMD decomposition method with the SARIMA seasonal time series method. This method is used to solve time series problems that have nonlinear characteristics and have seasonal patterns. The forecasting procedure with this method begins with a nonlinearity test of the data. The test results show that the data is nonlinear. Then the decomposition is carried out using the EMD method. The study indicates that the rainfall forecast in Bandung City for the next 12 months shows fluctuations where the highest rainfall occured in August 2020 and the lowest rainfall occured in January 2020. The forecasting results are still far from the original data but provide a MAPE value of 38,13307%, which means that the forecasting results are quite accurate. Keywords: Nonlinear, Seasonal Pattern, EMD, SARIMA, Rainfall.