Enhancing municipal pavement condition predictions using machine learning models

Pavement condition prediction is essential for effective street network maintenance. This paper examines the use of machine learning models to enhance prediction accuracy for municipal streets. Seventy-two models were developed to predict the 2022 pavement condition index using Linear Regression, Random Forest (RF) and Neural Network algorithms. The models were trained and tested on the 2018 and combined 2014–2018 datasets, incorporating 19 variables for residential and non-residential streets. Additional models were created for non-residential streets categorized by maintenance treatments. RF consistently outperformed other models, with the RF (A+D+S) model achieved the highest accuracy for residential streets, with a marginal improvement using the combined dataset. The RF (A+D+S+T) model performed best for non-residential streets and was also the most robust for pavement segments with surface levelling and milling & resurfacing. Weighted distress and Status2018 were the most significant variables. These models enhance accuracy in optimizing municipal street maintenance budgets.