A new construction algorithm of efficient radial basis function neural net classifier and its application to codes identification

In this paper we present a new simple algorithm to construct Radial Basis Function (RBF) neural net based classifier. This algorithm has the major advantage to require nothing else that the training set to work (no step learning, threshold or other parameters like in other methods). Despite its simplicity, we show, on many benchmark datasets, that this algorithm provides a robust and efficient classifier. These two properties make the proposed algorithm very attractive. We also describe an application of such built RBF classifier on data obtained in a project of buried codes identification. Finally, we compare the results with other new recognition techniques like fuzzy pattern recognition.