Author(s): Ujong JA and JG Egbe

Abstract: This study developed predictive models for pavement performance in Nigeria’s five geopolitical zones (North Central, North West, North East, South East, and South West) using artificial neural network (ANN) and multiple linear regression (MLR). The model aimed to identify the most significant factors influencing pavement performance and prediction functions. The result showed that the ANN and MLR both predicted pavement performance with the coefficient of determination (R2) 0.80785, 0.6780, 0.68997, 0.15602, and 0.93325 for the ANN model and mean square error (MSE) 1.8903, 2.4455, 1.9539, 1.690, 2.335, while the MLR model with R2 0.846913, 0.802864, 0.598531, 0.331692, 0.692314 for North Central, North East, Northwest, South West and South East respectively. However, the South West zone was poorly predicted by both models indicating unique characteristics or factors not captured by the model. This study highlights the importance of considering regional variations in pavement performance modelling. These findings can be a form of information for pavement and rehabilitation strategies, optimizing resource allocation, and improving road network performance in Nigeria.

DOI: 10.22271/27078388.2025.v6.i2a.49

Pages: 01-06 | Views: 503 | Downloads: 161

Download Full Article: Click Here