PERBANDINGAN AKURASI METODE AUTOREGRESSIVEINTEGRATED MOVING AVERAGE DAN SINGULAR SPECTRUMANALYSIS PADA PERAMALAN SUHU RATA-RATA HARIAN DI STASIUN METEOROLOGI AHMAD YANI
DOI:
https://doi.org/10.20527/h37agw97Keywords:
suhu udara, peramalan, ARIMA, SSAAbstract
Suhu udara di kota-kota besar dalam beberapa tahun terakhir mengalami peningkatan yang cukup ekstrim, sehingga diperlukan peramalan yang tepat untuk mengantisipasi terjadinya kerugian akibat suhu yang ekstrim. Penelitian ini bertujuan untuk membandingkan akurasi dari metode Autoregressive Integrated Moving Average (ARIMA) dan Singular Spectrum Analysis (SSA) dalam meramalkan suhu rata-rata harian di Stasiun Meteorologi Ahmad Yani. Data yang digunakan sebagai data training adalah data harian dari Januari 2022 hingga Desember 2022, sedangkan data harian Januari 2023 digunakan sebagai data testing. Pemodelan ARIMA diawali dengan menstasionerkan data, selanjutnya menentukan orde ARIMA berdasarkan plot autocorrelation function (ACF) dan partial autocorrelation function (PACF). Langkah-langkah pemodelan SSA yaitu embedding, singular value decomposition (SVD), grouping, dan diagonal averaging. Pemilihan model terbaik ditentukan berdasarkan nilai mean absolute percentage error (MAPE) terkecil. Hasil penelitian menunjukkan nilai MAPE data training dengan metode ARIMA(0,1,2) sebesar 2,0076%, sedangkan metode SSA sebesar 2,1777%. Nilai MAPE data testing dengan metode ARIMA(0,1,2) sebesar 3,2719%, sedangkan metode SSA sebesar 3,9464%. Berdasarkan hasil tersebut dapat disimpulkan bahwa metode ARIMA(0,1,2) lebih tepat digunakan untuk meramalkan data suhu rata-rata harian di Stasiun Meteorologi Ahmad Yani.
Air temperature in big cities in the last few years has increased quite extreme, so accurate forecasting is needed to anticipate losses due to extreme temperatures. This study aims to compare the accuracy of Autoregressive Integrated Moving Average (ARIMA) and Singular Spectrum Analysis (SSA) methods in predicting the daily average temperature at Ahmad Yani Meteorological Station. The data used as training data is daily data from January 2022 to December 2022, while the daily data for January 2023 is used as testing data. ARIMA modeling begins with stationary data, then determines the ARIMA order based on the autocorrelation function (ACF) and partial autocorrelation function (PACF) plots. SSA modeling steps are embedding, singular value decomposition (SVD), grouping, and diagonal averaging. The selection of the best model is determined based on the most minor mean absolute percentage error (MAPE). The results showed that value of MAPE training data using ARIMA(0,1,2) method was 2.0076%, while SSA method was 2.1777%. The MAPE value of testing data using ARIMA(0,1,2) method is 3.2719%, while SSA method is 3.9464%. Based on these results it can be concluded that the ARIMA(0,1,2) method is more appropriate for forecast daily average temperature data at Ahmad Yani Meteorological Station.
Downloads
References
Anwar, S. (2017). Peramalan Suhu Udara Jangka Pendek di Kota Banda Aceh dengan Metode Autoregressive Integrated Moving Average (ARIMA). Malikussaleh Journal of Mechanical Science and Technology, 5(1), 6–12.
Cristian, M. (2018). Average Monthly Temperature Forecast in Romania By Using Singular Spectrum Analysis. Academica Brancusi Publisher, 3, 48–57.
Ete, A. A. (2017). Forecasting The Amount Of Foreign Tourist Travelers to Indonesia by Entrance Guide Using Singular Spectrum Analysis and ARIMA (Thesis). Institut Teknologi Sepuluh Nopember, Surabaya.
Febrianti, T. K., Sulandari, W., & Pratiwi, H. (2021). Peramalan Curah Hujan di Kota Bandung Menggunakan Singular Spectrum Analysis. Jurnal Ilmiah Matematika, 8(2), 56–65.
Golyandina, N., & Korobeynikov, A. (2014). Basic Singular Spectrum Analysis and forecasting with R. Computational Statistics and Data Analysis, 71, 934–954. doi: https://doi.org/10.1016/j.csda.2013.04.009.
Hassani, H. (2007). Singular Spectrum Analysis: Methodology and Comparison. Journal of Data Science, 5(2), 239–257. doi: https://doi.org/10.6339/JDS.2007.05(2).396.
Khaeri, H., Yulian, E., & Darmawan, G. (2018). Penerapan Metode Singular Spectrum Analysis (SSA) Pada Peramalan Jumlah Penumpang Kereta Api di Indonesia Tahun 2017. Jurnal Euclid, 5(1), 8–20. doi: https://doi.org/10.33603/e.v5i1.496.
Lewis, C. D. (1982). International and Business Forecasting Methods. London: Butterworths.
Masyitoh, A. D., Zukhronah, E., & Slamet, I. (2022, March 24). Forecasting for Monthly Rainfall in Gondang Reservoir Lamongan Using Singular Spectrum Analysis (SSA) Method. doi: https://doi.org/10.1063/5.0075313.
Muis, S., & Setiyadi, D. (2020). Model Statistik Arima Dalam Meramal Pergerakan Harga Saham. Information System for Educators and Professionals, 4(2), 154–167.
Prasetyo, S., Hidayat, U., Haryanto, Y. D., & Riama, N. F. (2021). Variasi dan Trend Suhu Udara Permukaan di Pulau Jawa Tahun 1990-2019. Jurnal Geografi : Media Informasi Pengembangan Dan Profesi Kegeografian, 18(1), 60–68. doi: https://doi.org/10.15294/jg.v18i1.27622.
Pusat Database BMKG. (2015). Data Harian Temperatur Rata-Rata di Stasiun Meteorologi Ahmad Yani. Diakses pada 30 Januari 2023 dari http://dataonline.bmkg.go.id/data_iklim.
Rumagit, S. E., & Azhari. (2013). Prediksi Pemakaian Listrik Menggunakan Jaringan Syaraf Tiruan dan ARIMA Di Wilayah Sulluttenggo. Berkala Ilmiah MIPA, 23(2), 139–148.
Sakinah, A. M. (2018). Akurasi Peramalan Long Horizon dengan Singular Spectrum Analysis. Kubik: Jurnal Publikasi Ilmiah Matematika, 3(2), 93–99. doi: https://doi.org/10.15575/kubik.v3i2.4111.
Samidjo, J., & Suharso, Y. (2017). Memahami Pemanasan Global dan Perubahan Iklim. Journal IKIP Veteran Semarang, 24(2), 36–46.
Sodiqin, M. A., Sulandari, W., & Respatiwulan. (2021). The Application of Singular Spectrum Analysis Method in Forecasting The Number of Foreign Tourists Visit to Special Capital Region of Jakarta. Jurnal Riset Dan Aplikasi Matematika, 5(2), 92–102.
Solopos.com. (2022). Tak Hanya Jateng, Semarang Juga Termasuk Kota Terpanas di Indonesia. Dipetik pada 30 Januari 2023 dari https://www.solopos.com/tak-hanya-jateng-semarang-juga-termasuk-kota-terpanas-di-indonesia-1321732.
Wei, W. W. S. (2006). Time Series Analysis: Univariate and Multivariate Methods (2nd ed.). New York: Pearson Addison Wesley.
Downloads
Published
Issue
Section
License
Copyright (c) 2023 SENPIKA
This work is licensed under a Creative Commons Attribution 4.0 International License.
