Answers are grounded in this paper and cite their sources.
Reliable annual average daily traffic (AADT) estimates are essential for highway planning, pavement management, capacity assessment and infrastructure investment. This study develops and evaluates an integrated statistical and machine-learning framework for modelling AADT on five selected highways in South-Eastern Nigeria using demographic, economic, environmental and temporal predictors. The dataset comprises 100 state-year observations from 2001 to 2020 for the Umuahia–Ikot Ekpene, Onitsha–Owerri, Abakaliki–Ogoja, Enugu–Port Harcourt and Owerri–Port Harcourt highways. Predictors included passenger population density, gross domestic product (GDP), annual rainfall, pothole volume, year and highway effects. Pearson and Spearman correlations, multiple linear regression with HC3 heteroskedasticity-consistent inference, multicollinearity diagnostics and chronological machine-learning validation were applied. Mean pooled AADT was 1,313,110.42 PCU/day (SD = 569,491.41), with substantial spatial heterogeneity, ranging from 588,708.80 PCU/day on Abakaliki–Ogoja to 2,202,095.25 PCU/day on Enugu–Port Harcourt. Passenger population density showed the strongest Pearson association with AADT (r = 0.791), followed by GDP (r = 0.676), rainfall (r = 0.501), pothole volume (r = 0.307) and year (r = 0.141). The regression model explained 88.81% of in-sample variation (R² = 0.8881; adjusted R² = 0.8769). In chronological holdout testing for 2017–2020, Gradient Boosting achieved R² = 0.5338, RMSE = 396,916.31 PCU/day, MAE = 249,437.28 PCU/day and MAPE = 28.45%, outperforming the linear trend benchmark in RMSE. The findings demonstrate the value of combining interpretable statistical modelling with chronologically validated nonlinear prediction for AADT estimation in data-constrained highway systems.
No takes yet. Share an insight, caveat, or question.
Ugwuanyi et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: