MODELING MIXED GEOGRAPHICALLY WEIGHTED NEGATIVE BINOMIAL REGRESSION ON THE NUMBER OF TUBERCULOSIS CASES IN SOUTH SULAWESI

Abstract

Tuberculosis is an infectious disease caused by bacteria known as Mycrobacterium Tuberculosis, which is a problem in various regions, one of which is in the province of South Sulawesi which has experienced tuberculosis problems in recent years. Tuberculosis data in South Sulawesi shows overdispersion. This may be caused by the different geographical location of each region, so it is necessary to know the variables that affect tuberculosis cases. The overdispersion problem in the data can be overcome by using the Negative Binomial model. However, this model is only global while tuberculosis cases have different location characteristics. Therefore, a method is needed that can overcome overdispersion and consider the effects of spatial heterogeneity. Mixed Geographically Weighted Negative Binomial Regression (MGWNBR) is a model used for spatially heterogeneous discrete data that can overcome overdispersion in the data. The results of the study using MGWNBR show that the global variable that has a significant effect on the number of tuberculosis cases in all observed locations is the number of medical personnel, while the local variables that have a significant effect on the number of tuberculosis cases in some observed locations are the number of health facilities, population, and population density.

Access to Document

Other files and links

• Link to publication in Scopus
• Open Access Version Available

Fingerprint