Combining use of rough set and artificial neural networks in doweled-pavement-performance modeling - A hybrid approach

This paper explores the potential applications of rough set theory and neural networks in concrete-faulting-performance modeling. The rough set can be used to reduce the dimension of the original pavement database in terms of attributes and rows. The reduced table can then be used as an input into the neural network. Since there are no universal rules for neural network construction, this approach is very important. Constructive and destructive methods have been used in the construction of neural network architecture. Although these methods are appropriate, they do not have a strong scientific proof. The key characteristics of the proposed method is that the new decision table created by using the rough set analysis will free the neural network paradigm from redundancy. A simple example for faulting performance in concrete pavement is presented.