NeuralNetType

NeuralNetType

NeuralNetType is a property of a NeuralNet object. It specifies the topology of the neural net. It can be: Layer for a feedforward network where the neurons in each layer are only connected to the neurons in the next layer. Shortcut is a feedforward net where the neurons are connected to the neurons in all the following layers. Layer is the default type. Changing this property will reset the weights in the net.

Layer

Layer is a value of the Neural Net property NeuralNetType. The Layer net type is a feedforward network where the neurons in each layer are only connected to the neurons in the next layer. The ConnectionRate can specify the number of connections created. A fully connected net of the Layer NeuralNetType is the default.

Shortcut

Shortcut is a value of the Neural Net property NeuralNetType. The Shortcut net type is a feedforward net where the neurons are connected to the neurons in all the following layers. It can only be created fully connected i.e. the ConnectionRate is fixed at 1.

Created by freegoldbar  (September 16, 2004)