Estimating the Parameters of a Robust Geographically Weighted Regression Model in Gross Regional Domestics Product in East Java

  • Bayutama Isnaini Department of Statistics, IPB University, Bogor, 16680, Indonesia
  • Utami Dyah Syafitri Department of Statistics, IPB University, Bogor, 16680, Indonesia
  • Muhammad Nur Aidi Department of Statistics, IPB University, Bogor, 16680, Indonesia
Keywords: Outliers, M-estimator RGWR, S-estimator RGWR


Geographically weighted regression (GWR) is a regression parameter estimation method that accommodates location elements. Estimates of regression parameters have problems when there are outliers in the modelled data, including data based on location. This problem can be handled by a robust method of outliers, the robust GWR method (RGWR). M-estimator and S-estimator have high efficiency and high breakdown points. This study aimed to determine the best regression parameter estimation model on gross regional domestic product (GRDP) data in East Java Province in 2015, which is indicated to have various value based on the characteristic of regency/city. The city of Surabaya has very different characteristics from other regions and is detected as outliers based on a GWR model error plot, so RGWR with M-estimator and S-estimator are used. The mean absolute deviation (MAD) ​​show that the best model for data in this study is the RGWR with M-estimator.


. H. Zhang and C. Mei. “Local least absolute deviation estimation of spatially varying coefficient models: robust geographically weighted regression approaches.” International Journal of Geographical Information Science. vol. 25(9), pp. 1467-1489, 2011.

. M.J.S. Windle, et al. “Exploring spatial non-stationarity of fisheries survey data using geographically weighted regression (GWR): an example from the Northwest Atlantic.” ICES Journal of Marine Science. vol. 67, pp.145–154, 2010.

. A.S. Fotheringham. et al. Geographically Weighted Regression: The Analysis of Spatially Varying Relationship. Chichester, UK: John Wiley & Sons Ltd, 2002, pp. 27-64.

. Z. Mahmood. and Salahuddin. “Resampling Method for the Data Adaptive Choice of Tuning Constant in Robust Regression.” Pakistan Journal of Statistics. vol. 31(3), pp. 281-294, 2015.

. P.J. Rousseeuw and V.J. Yohai. “Robust Regression by Means of S-Estimators, Robust and Nonlinear Time Series Analysis.” Lecture Notes in Statistics. vol. 26, pp. 256–272, 1984.

. V. Verard and C. Croux. “Robust regression in stata.” The Stata Journal. vol. 9(3), pp. 439-453, 2009.

. P. Harris. et al. “Robust geographically weighted regression: a technique for quantifying spatial relationships between freshwater acidification critical loads and catchment attributes.” Annals of the Association of American Geographers. vol. 100(2), pp. 286 – 306, 2010.

. R. Afifah. “Robust geographically weighted regression with least absolute deviation method in case of poverty in Java Island.” American Institute of Physics Conference Proceedings. vol. 1827(1), pp. 020023, 2017.

. I. C. Nurhayati et al. “Robust geographically weighted regression with least absolute deviation (case study: the percentage of diarrhea occurrence in semarang 2015).” Journal of Physics: Cofference Series. vol. 1217, pp. 012099, 2019.

. Badan Pusat Statistik (BPS). Provinsi Jawa Timur dalam Angka 2017: Jawa Timur, ID, 2017.

. M. Solihin. “Pengembangan regresi terboboti geografis dan temporal menggunakan interaksi jarak spasial temporal studi kasus pertumbuhan ekonomi di jawa tengah tahun 2011-2015.” M.Sc. thesis, IPB University, Indonesia, 2017.

. I. Yulita. “Pemodelan regresi komponen utama dan LASSO terboboti geografis (global dan lokal) (studi kasus data produk domestik regional bruto (PDRB)) pada 113 kabupaten/kota di pulau jawa.” M.Sc. thesis, Indonesia, 2016.

Y. Susanti et al. “M estimation, S estimation, and MM estimation in robust regression.” International Journal of Pure and Applied Mathematics. vol. 91(3), pp. 349-360, 2014.