Single layer feedforward networks construction based on orthogonal least square and particle swarm optimization

Abstract

According to the simplicity and universal approximation capability, single layer feedforward networks (SLFN) are widely used in classification and regression problems. The paper presents a new OLSPSO constructive algorithm based on Orthogonal Least Square (OLS) method and Particle Swarm Optimization (PSO) algorithm. Instead of evaluating the orthogonal components of each neuron as the conventional OLS method, a new recursive formulation is derived. Then based on the new evaluation of each neuron’s contribution, the PSO algorithm is used to seek the optimal parameters of the new neuron in continuous space. The proposed algorithm is experimented on some practical regression problems and compared with other constructive algorithms. Results show that proposed OLS-PSO algorithm could achieve a compact SLFN with good generalization ability.