Robust spatial Durbin model for dengue fever cases in Indonesia

Abstract

Spatial regression analysis is a method employed for data exhibiting spatial dependencies. The Spatial Durbin Model (SDM) is one such method that accounts for spatial effects in both the dependent and independent variables. However, the presence of spatial outliers can compromise the accuracy of SDM predictions. To address this issue, a robust method is needed, namely the Robust Spatial Durbin Model (RSDM). This study applies the RSDM to model the factors influencing the spread of dengue haemorrhagic fever cases across Indonesia and to identify the superior modeling approach. The results indicate that the RSDM outperforms the standard SDM, evidenced by a higher Adjusted R² and a lower Mean Squared Error (MSE). Key factors identified as significantly influencing dengue haemorrhagic fever cases are population density, the number of doctors in healthcare facilities, and the percentage of the population covered by health insurance.