Input Selection Using Binary Particle Swarm Optimization

Nowadays, multi-layer feed forward networks are often used for modeling complex relationships between the data sets. And if we can choose only the important data from the training sets, it will make the networks less size and can save more time. Because we realize in this point, this paper provides procedure of feature selection to train the neural networks using binary particle swarm optimization. It also introduces the suitable function for the binary particle swarm optimization technique by changing concept in part of member value adjustment function for each particle.