Model Geographically Weighted Weibull Regression Pada Indikator Pencemaran Air COD di Daerah Aliran Sungai Mahakam Kalimantan Timur

Primadigna, Ullimaz Sam
Suyitno, Suyitno
Siringoringo, Meiliyani

Eksponensial
Vol. 13 No. 2 (2022)

The Geographically Weighted Weibull Regression (GWWR) model is a Weibull regression model applied to spatial data. Parameter estimation is carried out at each observation location using spatial weighting. This study aimed to determine the GWWR model on the Chemical Oxygen Demand (COD) water pollution indicator data and to obtain the factors that influence COD in the Mahakam watershed. The parameter estimation method was Maximum Likelihood Estimation (MLE). Spatial weighting in parameter estimation has been determined using the adaptive tricube weighting function and the criteria for determining the optimum bandwidth was Generalized Cross-Validation (GCV). The research sample was 20 location points of the Mahakam river determined by the Environmental Department of East Kalimantan Province. The results showed that the factors that influence COD locally was temperature, while the factors that influence globally were temperature, Total Suspended Solids (TSS), and Fecal Coli.