The Application of Parametric and Non-Parametric Models in Analysing Urban Expansion in the South-South Nigeria

Abstract

This paper applies parametric and non-parametric models in analysing urban growth patterns in the South-South region of Nigeria. A mixed-methods framework integrating geospatial analysis, statistical techniques, and machine learning was adopted. Landsat imagery (1993, 2003, 2013, and 2023) was processed through supervised Land Use/Land Cover classification, and an Urban Expansion Index (UEI) was computed to evaluate growth efficiency. Physical factors such as elevation and slope, infrastructural variables including proximity to highways, central business districts, and ports, and socio-economic attributes, such as access to markets and population density, were derived from spatial analysis and complemented with questionnaire data. Parametric models (linear and logistic regression) were compared with non-parametric approaches (Multi-Layer Perceptron Neural Networks, Random Forest, and Geographically Weighted Regression). Findings reveal significant urban sprawl across all cities studied, with Benin City’s urban footprint rising from 13.7% in 1993 to 29.2% in 2023, and Port Harcourt’s from 10.5% to 36.8%. Key drivers included population growth, industrialisation, and proximity to infrastructural hubs, while elevation constrained development. Non-parametric models outperformed parametric ones, highlighting complex non-linear dynamics. The study details the strengths and limitations of each modeling approach and their effectiveness in capturing the complexities of urban expansion. The implications of these findings for urban planning and future research are discussed.