Study of Clustering Time Series Forecasting Model for Provincial Grouping in Indonesia Based on Rice Price
Kajian Model Peramalan Clustering Time Series untuk Penggerombolan Provinsi Indonesia berdasarkan Harga Beras
DOI:
https://doi.org/10.29244/ijsa.v6i1p50-62Keywords:
ARIMA, clustering time series, correlation distance, rice priceAbstract
Most indonesians consume rice as the main staple. The high low price of rice has an impact on farmers and communities, especially those who cannot afford it. Rice price forecasting is one of the important information to be considered for future rice prices. The data used is secondary data sourced from bps publication, Rural Consumer Price Statistics: Food Group, from January 2008 to December 2019 for 32 provinces in Indonesia. Time series modeling and forecasting is usually done on a single variable using ARIMA. however, modeling becomes inefficient if there are many variables, so clustering time series analysis is performed using correlation distance with the clustering method of average linkage hierarchy. Cluster level ARIMA modeling with 4 clusters provides high efficiency because only by doing 4 times modeling results in accuracy values not much different from individual level modeling. the results obtained by individual-level ARIMA Modeling resulted in an average MAPE of 3.36%, while cluster-level ARIMA modeling with 4 clusters resulted in an average MAPE value of 4.27%, with a second MAPE difference of -0.91%. Formally conducted z test, the results obtained there is no difference between individual-level MAPE and cluster-level MAPE. This means that cluster-level modeling is relatively good and representative.
Downloads
References
[BPS] Badan Pusat Statistik. 2018. Statistik Harga konsumen Perdesaan Kelompok Makanan . Jakarta [ID]: Badan Pusat Statistik.
Cryer JD, Chan KS. 2008. Time Series Analysis with Applications in R 2th Edition.Ed Ke-2. New York [US]: Springer.
D’Urso P, Maharaj EA. 2009. Autocorrelation-based fuzzy clustering of time series. Fuzzy Sets and Systems 160[24]: 3565-3589.
Golay X, Kollias S, Stoll G, Meier D, Valavanis A, Boesiger P. 2005. A new correlation based fuzzy logic clustering algorithm for FMRI. Magnetic Resonance in Medicine. 40[2]: 249–260.
Kalkstein LS, Tan G, dan Skindlov JA. 1987. An evaluation of three clustering procedures for use in synoptic climatological classification. Journal of climate and applied meteorology 26[2], 717-730.
Kaufman L, Rousseeuw PJ. 1990. Finding Groups in Data: An Introduction to Cluster Analysis. New Jersey [US]: John Wiley and Sons Inc.
[Kementan RI] Kementerian Pertanian Republik Indonesia. 2019. 10 Besar Provinsi Penghasil Beras. [Internet]. Tersedia pada : https://www.pertanian.go.id/home/?show=news&act=view&id=4425.
Kleiber C, Zeileis A. 2008. Applied Econometrics with R. New York [US]: Springer Verlag.
Kumar M. 2005. Combining forecasting using clustering. Rutcor research report [RRR] 40-2005.
Maharaj EA, Brett AI. 1999. Forecasting time series from clusters. Departemen of econometrics and business statistics Monash University ISSN 1440-771X ISBN 0 7326 1064 8.
Mardianto, Gunawan, Sugiarto dan Rochman . 2020. Perbandingan Peramalan Harga Beras menggunakan Metode ARIMA pada Amazon Forecast dan Sagemaker. Jurnal RESTI. Vol. 4 No. 3, 537 - 543.
Montero P, Vilar JA. 2014. Tsclust: An R Package for Time Series Clustering. Journal of Statistical Software, November 2014. 62[1]: 1-43.
Montgomery CD, Jennings CL, Kulahci M. 2008. Introduction to Time Series Analysis and Forecasting. Third Ed. New Jersey [US]: John Wiley $ Sons, Inc.
Munthe AD. 2019. Penerapan Clustering Time Series untuk menggerombolkan provinsi di Indonesia berdasarkan Nilai Produksi Padi. Jurnal Litbang Sukowati. 2[2]:1-11.
Novidianto dan Dani. 2020. Analisis Klaster Kasus Aktif Covid-19 menurut Provinsi di Indonesia berdasarkan Data Deret Waktu. Jurnal Aplikasi Statistika dan Komputasi Statistik. 12[2] ISSN 2086-4132.
Rencher AC. 2002. Methods of Multivariate Analysis 2nd ed. New York [US]: A John Wiley & Sons, Inc.
Sekarsari UW dan Ngatini. 2019. Pengembangan Pemodelan Harga Beras di Wilayah Indonesia Bagian Barat dengan Pendekatan Clustering Time Series. Journal of Mathematics and Its Applications.
Sokal RR, Rohlf FJ. 1962. The Comparison of Dendograms by Objective Methods. International Association for Plant Taxonomy [IAPT], February 1962.11[2]: 33-40.
Siswanto, Sinaga dan Harianto. 2018. Dampak Kebijakan Perberasan pada Pasar Beras dan Kesejahteraan Produsen dan Konsumen Beras di Indonesia. Jurnal Ilmu Pertanian Indonesia [JIPI]. Vol. 23 [2]: 93-100.
Utami B. 2018. Model VARX untuk Peramalan Inflasi Menurut Sub Kelompok Komoditi di Jakarta dengan Pendekatan TSClust sebagai Preprocessing[Tesis]. Bogor [ID] : Institut Pertanian Bogor.
Wei WWS. 2006. Time Series Analysis, Univariate and Multivariate Methods, Second Edition. New York [US]: Pearson Education, Inc.
