PEMODELAN POTENSI PRODUKSI KUBIS (BRASSICA OLERACEA) DI KABUPATEN MALANG MENGGUNAKAN GEOGRAPHICALLY WEIGHTED REGRESSION (GWR) DENGAN PENDEKATAN PEMBOBOTAN BI-SQUARE EKSPONENSIAL DAN GAUSSIAN
DOI:
https://doi.org/10.54314/cfy3qa57Keywords:
Geographically Weighted Regression, Weighting Function, CabbageAbstract
Abstract: Cabbage is one of the horticultural crops with the largest production in Indonesia. Cabbage in Malang Regency is one of the main commodities widely planted by farmers. Geographically Weighted Regression (GWR) is a statistical method used to analyze spatial influences with a point approach that shows that between locations have a tendency towards different characteristics in representing a condition. This model is applied in the agricultural sector to estimate cabbage production that varies between regions. The purpose of this study is to determine the best weighting to form a model and variables that influence cabbage production in Malang Regency, using the GWR model with fixed kernel weighting Bi-square, Exponential, and Gaussian. The best criteria are based on the coefficient of determination (R2) and the Akaike Information Criterion (AIC) statistics. Secondary data used are sourced from the Food Crops, Horticulture and Plantation Service of Malang Regency, with predictor variables including land area, elevation, and rainfall. Based on the R2 and AIC values, the analysis results indicate that the GWR model with an exponential weighting function is the best model for cabbage production, and land area is the variable that influences cabbage production in Malang Regency.
Keywords: Geographically Weighted Regression, weighting function, cabbage
Abstrak. Kubis menjadi salah satu tanaman hortikultura yang termasuk dalam produksi terbesar di Indonesia. Kubis di Kabupaten Malang merupakan salah satu komoditi utama yang banyak ditanam oleh petani. Geographically Weighted Regression (GWR) merupakan salah satu metode statistika yang digunakan untuk menganalisis pengaruh spasial dengan pendekatan titik yang menunjukkan bahwa antar lokasi memiliki kecenderungan terhadap karakteristik yang berbeda dalam merepresentasikan suatu kondisi. Model ini diterapkan pada bidang pertanian untuk menduga produksi kubis yang bervariasi antar daerah. Tujuan penelitian adalah untuk menentukan pembobot yang terbaik untuk membentuk model dan peubah yang berpengaruh terhadap produksi kubis di Kabupaten Malang, menggunakan model GWR dengan pembobot fixed kernel Bi-square, Eksponensial dan Gaussian. Kriteria terbaik didasarkan pada koefisien determinasi (R2) dan statistik Akaike Information Criterion (AIC). Data sekunder yang digunakan bersumber dari Dinas Tanaman Pangan Hortikultura dan Perkebunan Kabupaten Malang dengan peubah prediktor meliputi luas lahan, elevasi dan curah hujan. Berdasarkan nilai R2 dan AIC hasil analisis menunjukkan bahwa model GWR dengan fungsi pembobot eksponensial merupakan model terbaik untuk produksi kubis dan luas lahan adalah peubah yang berpengaruh terhadap produksi kubis di Kabupaten Malang.
Kata Kunci : Geographically Weighted Regression, fungsi pembobot, kubis
Downloads
References
Badan Pusat Statistik, 2015. Statistik Pertanian Hortikultura Jawa Tengah 2012-2014
Badan Pusat Statistik. 2025. Statistik Hortikultura 2024. Nomor Katalog 5204003; Nomor Publikasi 05100.25005. Jakarta: BPS.
Fahrmeir, L., Kneib, T., Lang, S., & Marx, B. D. (2021). Regression: Models, methods and applications (2nd ed.). Springer. https://doi.org/10.1007/978-3-662-63882-8.
Ahmad, M., & Sari, L. P. (2021). Spatial modeling approaches: Point and area-based methods in agricultural data analysis. Journal of Geospatial Science, 15(2), 134–147.
Putra, R. A., & Dewi, S. K. (2023). Application of geographically weighted regression (GWR) with fixed bi-square kernel in horticultural crop production analysis. Indonesian Journal of Agricultural Science, 19(1), 45–56.
Ramadhan, T., Nugroho, A., & Wibowo, F. (2020). Spatial variation in agricultural activities and farmer access to facilities using geographically weighted regression in South Sumatra. Journal of Environmental and Agricultural Studies, 8(3), 221–233.
Wang, Y., & Li, Z. (2022). The impact of weight matrix selection in geographically weighted regression models: A review and case study. Spatial Statistics, 50, 100600.
Anismuslim, M., Pramoedyo, H., Andarini, S., & Sudarto. (2023). Modeling of Risk Factors of Childhood Stunting Cases in Malang Regency using Geographically Weighted Regression (GWR). The Open Public Health Journal, 16(1).
Comber, A., Brunsdon, C., Charlton, M., Dong, G., Harris, R., Lu, B., Lü, Y., Murakami, D., & Wang, Y. (2020). The GWR route map: a guide to the informed application of Geographically Weighted Regression. arXiv preprint.
Idris, M. A., & Aidi, M. N. (2024). Pemodelan Geographically Weighted Regression (GWR) pada Prevalensi Severely Stunting di Indonesia Tahun 2023. Journal of Mathematics: Theory and Applications, 7(1).
Laome, L., Budiantara, I. N., & Ratnasari, V. (2024). Construction of a geographically weighted nonparametric regression model fit test. MethodsX, 12, 102536.
Noerul Hanin, I., Meilandra, I., Debataraja, N. N., & Pertiwi, R. (2024). Application of Geographically Weighted Regression for Modeling the Poverty Cases in Kalimantan, Indonesia. Jurnal Statistika dan Aplikasinya, 8(1), 1–16.
Yusnita, D., Sari, I. P., & Zulfikri. (2022). Analisis Faktor-Faktor yang Mempengaruhi Produksi Kubis di Kabupaten Wonosobo. Jurnal Agribisnis Indonesia, 10(1), 45–52.
