ISSN: 2966-0599
v.2, n.1, 2025 (Janeiro)
METADADOS
DOI: 10.69720/2966-0599.2025.00035
Author 1: Dodo Boubakar
Biography: Department of Statistics, Usman Dan Fodiyo University Sokoto, Nigeria
E-mail: dboubakar6@gmail.com
==========================================
Author 2: Umar Usman
Biography: Department of Statistics, Usman Dan Fodiyo University Sokoto, Nigeria
E-mail: uusman07@gmail.com
==========================================
Author 3: B K asare
Biography: Department of Statistics, Usman Dan Fodiyo University Sokoto, Nigeria
E-mail: Mponua2004@yahoo.com
==========================================
Author 4: Y M Ahijjo4
Biography: Department of Physics, Usman Dan Fodiyo University Sokoto, Nigeria
E-mail: yahijjomusa@gmail.com
==========================================
Author 5: A. M. Ndatsu
Biography: Department of Statistics, Usman Dan Fodiyo University Sokoto, Nigeria
E-mail: da6thabdul@gmail.com
==========================================
ABSTRACT: This study investigates crime patterns using regression kriging. The methodology involves determining variogram models for various crime types and fitting multiple regression kriging models. Most variogram models were Gaussian, with some exhibiting smooth transition estimates or spherical structures. The analysis revealed that the majority of crime types exhibit spatial correlation, with range values greater than zero. In the regression kriging models, population size and literacy rate emerged as the most significant predictors. Additionally, the variograms of the kriging residuals indicated trends for certain crime types. However, for most crimes, the residuals displayed spatial autocorrelation, suggesting that the models are suitable for predictive purposes.
Keywords: Variogram, Rgression kriging
