Modified Cascade-2 Algorithm with Adaptive Slope Sigmoidal Function

Abstract

Cascade-2 algorithm is a variant of well-known cascade-correlation algorithm that is widely investigated  constructive training algorithm for designing cascade feed forward neural networks. This paper proposes a  modified Cascade-2 algorithm with adaptive slope sigmoid function (MC2AASF). The algorithm emphasizes on  architectural adaptation and functional adaptation during learning. This algorithm is a constructive approach of  designing cascade architecture. To achieve functional adaptation, the slope of the sigmoid function is adapted  during training. One simple variant is derived from MC2AASF is where the slope parameter of sigmoid function  used at the hidden layers’ nodes is fixed to unity. Both the variants are compared to each other on five function  approximation tasks. Simulation results show that adaptive slope sigmoid function presents several advantages  over standard fixed shape sigmoid function, resulting in increasing flexibility, smoother learning, better  generalization performance and better convergence