An incremental learning algorithm for function approximation

Abstract: This paper presents an incremental learning algorithm for feed-forward neural networks used as approximators of real world data. This algorithm allows neural networks of limited size to be obtained, providing better performances. The algorithm is compared to two of the main incremental algorithms (Dunkin and cascade correlation) in the respective contexts of synthetic data and of real data consisting of radiation doses in homogeneous environments. [Copyright &y& Elsevier]