Since feed-forward artificial neural networks (FFANNs) are the most widely used models to solve real-life problems, many studies have focused on improving their learning performances by changing the network architecture ...