Use of Recursive Partitioning to Predict National Bridge Inventory Condition Ratings from National Bridge Elements Condition Data
In the United States, National Bridge Inventory (NBI) condition ratings, since the 1970s, and AASHTO’s commonly recognized (CoRe) element condition data, since the 1990s, have provided two major sources of bridge condition data. Although these separate systems of condition assessment had their individual uses, comparing the two, and mapping one from the other had uses for both state and federal agencies and the bridge management community. Alternative methods for this mapping have been proposed in the literature with varying predictive accuracy. With the publication of the new AASHTO Manual for Bridge Element Inspection in 2013, national bridge elements (NBEs) replace the CoRe element condition data as the comparable condition data for the NBI condition ratings. This paper investigates the use of the recursive partitioning method to develop classification trees that predict NBI condition ratings from NBE condition data. On the basis of data from a 2016 submission and 12 transportation agencies, classification trees were developed that presented the most likely NBI condition ratings for a set of logical conditions based on the relative element quantities and the percentage of element quantities in the condition states. The predictive accuracies for the trees are sufficient, and the percentages of exact matches and matches within one error term are better than other studies in the literature. Although the trees can be improved in the future with the availability of more NBE data submissions, the study presented preliminary decision trees with sufficient predictive accuracy that could be adopted by transportation agencies for a variety of bridge management functions.
Mechanical and Civil Engineering
Transportation Research Record: Journal of the Transportation Research Board
Bektaş, B. A. (2017). Use of Recursive Partitioning to Predict National Bridge Inventory Condition Ratings from National Bridge Elements Condition Data. Transportation Research Record: Journal of the Transportation Research Board, 2612(1), 29–38. https://doi.org/10.3141/2612-04
Link to Publisher Version (DOI)
Publisher's Copyright and Source
Copyright © SAGE
Article published in Transportation Research Record: Journal of the Transportation Research Board, volume 2612, issue 1, 2017, pages 29-28.