|
Node Freezing is coming in version 10. Nodes are frozen before or during the
learning process. Node freezes are set to any
level. The freeze levels are then reduced at any rate
while the network is learning. Nodes do not take part in the
learning process while they are frozen. When a nodes freeze level drops to
zero the node is added back to the network. Node freezing finds the simplest, most efficient network and determines which
nodes and inputs are the most important and which are not needed.
The whole process is automatic.
Copyright
©
2002-2008 Neural Planner Software Ltd
Last Updated: 8th
May 2008 |
