It is found that there is a tradeoff between the performance of a
network and the time consumed. Generally, the performance of a
network is found to increase with suitable increase in the number
of samples, epochs (learning time), and the number of hidden layer
nodes. Meanwhile, increase of these parameters also increases the
consumed time. It may be said that learning rate (a) and momentum
term (g), in addition to hidden layer nodes, epochs, and number
of samples, also greatly influence the network to provide good
generalization [69].
It is found that there is a tradeoff between the performance of anetwork and the time consumed. Generally, the performance of anetwork is found to increase with suitable increase in the numberof samples, epochs (learning time), and the number of hidden layernodes. Meanwhile, increase of these parameters also increases theconsumed time. It may be said that learning rate (a) and momentumterm (g), in addition to hidden layer nodes, epochs, and numberof samples, also greatly influence the network to provide goodgeneralization [69].
การแปล กรุณารอสักครู่..
