Improved Classification Algorithm for The Counter Propagation Network

Abstract

In this paper the classification process of the Counter Propagation Network (CPN) is investigated. In the first part, the structure and the learning algorithm of the CPN is detailed. The second section gives an overview of general classification process. The paper proposes a modification of the original CPN classification algorithm to reduce the misclassification error in the region of uncertain decision. The accuracy of the proposed algorithm is tested with a case study.