Keywords: bridge performance, deterioration, market basket analysis, association rule mining, maintenance decision-making.

Maintenance of highway bridges plays an important role to assure the adequate reliability and desirable service of highway networks. The repercussions of the deterioration or the damage of highway bridges during their service life presents their own unique concerns and safety challenges. The primary goal of a bridge maintenance management is to assist bridge managers in making consistent and cost-effective decisions for bridge maintenance and rehabilitation. Such decisions consist of determining optimal budget allocations, establishing policies priorities, and the frequency of maintenance, and assessing conditions and residual lives of bridge components.

A degree-extent-relevancy-urgency (DERU) evaluation approach is currently used by the bridge administration units for rating bridge conditions in Taiwan [1]. Twenty components including deck, girder, pier, pavement, sidewalk, cap beam, bearing, joint, abutment, retaining wall, drainage, foundation and others are identified in the DERU method for bridge inspection in practice. Regularly, the bridges are visually inspected once every two years. Numerous highway bridges have been discovered that they were partially or seriously deteriorated or damaged before reaching their service lives by time, traffic loading, environmental factors and maintenance activities in Taiwan. Failure or deterioration of constructed bridge structures has resulted in time and cost overrun, injuries, and fatalities. Since bridge components are usually subject to similar environmental conditions, their deterioration is usually correlated. Discovering meaningful patterns and association rules in bridge inspection data is useful to establish maintenance policies. It is also of fundamental importance to utilize useful techniques to assist the bridge managers to diagnose potential bridge failure origins, which may lead to catastrophic consequences. An association rule expresses an association between items or sets of items. By induction of the association rule, sets of data instances that frequently appear together can be found. Since association rules are useful and easy to understand, there have been many applications, for example [2,3,4].

Market basket analysis (MBA), also known as association rule mining, is an effective data mining method used to determine which items are most frequently occurred jointly. This paper applies market basket analysis to discover deterioration patterns by extracting associations or co-occurrences from the inspection data of 112 highway bridges in Taiwan. The results demonstrate the capability of this approach, which can help the bridge managers to better identify the interdependencies of deteriorated bridge components and make a proper bridge maintenance policy.

1

N.-F. Pan, "Forecasting bridge deck conditions using fuzzy regression analysis", Journal of the Chinese Institute of Engineers, 30(4), 593-604, 2007.

2

D.-R. Liu, Y.-Y. Shih, "Integrating AHP and data mining for product recommendation based on customer lifetime value", Information & Management, 42(3), 387-400, 2005. doi:10.1016/j.im.2004.01.008

3

K. Koperski, J. Han, "Discovery of spatial association rules in geographic information database", Proceedings of the 4th international Symposium on Large Spatial Databases (SSD 95), Portland, Maine, 47-66, 1995.

4

X.-B. Li, "A scalable decision tree system and its application in pattern recognition and intrusion detection", Decision Support Systems, 41 (1), 112-130, 2005. doi:10.1016/j.dss.2004.06.016

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £95 +P&P)