Skip to Main Content
A designing method for counter-propagation neural networks based on rough set theory is presented in this paper. Counter-propagation networks has been applied to various fields because of its topological construction closed to the mankindpsilas brain, while rough set theory has a powerful capability for qualitative analysis. By combining those advantages of the two theories, we can construct a kind of neural networks with good understandability, simple computation and exact accuracy. In this paper, the key of the algorithm is that the input amples are simplified and classified by using rough set theory before trained.