← All articles

Photo · Gordon More

Data-Driven Approach to State of Good Repair: Predicting Rolling Stock Service Life with Machine Learning for State of Good Repair Backlog Reduction and Long-Range Replacement Cost Estimation in Small Urban and Rural Transit Systems

Dilip Mistry, Jill Hough · 2024 · Transportation Research Record Journal of the Transportation Research Board

Summary. This paper develops a machine learning model to predict when transit vehicles need replacement in small urban and rural U.S. transit systems. Using historical data from retired vehicles, the model applies random forest and gradient boosting techniques to estimate service life, identify maintenance backlogs, and forecast replacement costs. The tool helps transit agencies maintain vehicles in good repair, reduce backlogs, and make better funding decisions for asset management.

Read the original

Cite this article

Mistry, D., & Hough, J.. (2024). Data-Driven Approach to State of Good Repair: Predicting Rolling Stock Service Life with Machine Learning for State of Good Repair Backlog Reduction and Long-Range Replacement Cost Estimation in Small Urban and Rural Transit Systems. Transportation Research Record Journal of the Transportation Research Board. https://doi.org/10.1177/03611981241235197

Details

DOI
10.1177/03611981241235197
Countries
United States
Regions
North America
Categories
transportation, rural-data-and-definitions
Added
2026-04-28