Rough Set Theory in Pavement Maintenance Decision

The decision tree models based on pavement experts' experiences are usually used to derive pavement maintenance decisions. Because experiences are very difficult to obtain completely, some importance condition attributes in decision tree model are easily to be lost. With the prevailing of the computer, Case Base Reasoning (CBR) or Data Mining approach are used to process a large number of historical pavement maintenance records. But, the shortcoming is all condition attributes have to be included in the analyses. Rough Set Theory was proposed in 1982. Rough Set Theory can reduce the condition attributes and decision rules without changing the accuracy. The provincial and county roads located in north Taiwan are used as the empirical study. With the same accuracy, Rough Set Theory can reduce thirteen condition attributes to eight. Thus, RST is suitable for pavement maintenance decisions.