In recent years, pavement engineering has gradually shifted from new construction work to pavement maintenance and management. Since pavement engineers of the Taipei City Government change frequently, objective data is used to make decisions pertaining to road maintenance in Taipei City instead of relying on engineers' experience. In this study, three methods (ID3, C5.0 and SVM) have been chosen to test for use in the decision-making process related to road maintenance of Taipei City. The results show the correct classification rates of the decision trees are 76.67% (C5.0), 64.52% (ID3), and 66.67% (SVM). The decision tree of C5.0 was compared with engineer's experience, with 70% conformity between these two methods. Although the accuracy of the classification could be further improved, the decision tree of C5.0 could be used for pavement maintenance instead of human judgment.

Included in

Engineering Commons