TwoLayerNetwork Class Reference

Two layered networks are very commonly used. More...

#include <TwoLayerNetwork.h>

Inheritance diagram for TwoLayerNetwork:

[legend]Collaboration diagram for TwoLayerNetwork:

[legend]List of all members.

Public Member Functions
	TwoLayerNetwork (uint inputs, uint hidden, uint outputs)
	Creates a two layer network.
	TwoLayerNetwork (const char *filename)
	Loads a two-layer network from a file.
virtual void	connect2in (int input, int hidden)
	Connects an input and a hidden neuron with a random weight.
virtual void	connect2in (int input, int hidden, real weight)
	Connects an input and a hidden neuron with the given weight.
virtual void	connect2out (int hidden, int output)
	Connects a hidden and an output neuron with a random weight.
virtual void	connect2out (int hidden, int output, real weight)
	Connects a hidden and an output neuron with the given weight.
virtual void	connectAll ()
	Completely connects the network.
virtual void	addLayer (int size)
	Overrides MultiLayerNetwork::addLayer() so that it cannot be done.
virtual const char *	getClassName ()
	Returns "TwoLayerNetwork".

---

Detailed Description

Two layered networks are very commonly used.

This is basically a multi-layer perceptron network with only two layers - one hidden and one output (the input is not counted as a layer). This class basically derives from a MultiLayerNetwork and adds functionality that make it easier to use if the network you're dealing with has only 2 layers. There is nothing you can do with this class that you can't do with MultiLayerNetwork, just that this may be easier to use.

---

Constructor & Destructor Documentation

TwoLayerNetwork (  uint  inputs, uint  hidden, uint  outputs )

Creates a two layer network. Parameters: inputs  The number of inputs taken in by the network hidden  The number of hidden neurons in the network outputs  The size of the output vector (number of outputs given by the network)

TwoLayerNetwork (  const char *  filename  )

Loads a two-layer network from a file. The file is exactly the same as a MultiLayerNetwork file, and can be loaded there as well. Exceptions: Exception  if the file read does not correspond to a two layer network

---

Member Function Documentation

virtual void addLayer (  int  size  )  [virtual]

Overrides MultiLayerNetwork::addLayer() so that it cannot be done. Exceptions: Exception  Because the number of layers of this class is fixed, so you shouldn't be allowed to add a layer Reimplemented from MultiLayerNetwork.

virtual void connect2in (  int  input, int  hidden, real  weight )  [virtual]

Connects an input and a hidden neuron with the given weight. Parameters: input  The index of the input neuron hidden  The index of the hidden neuron in the hidden layer weight  The weight of the connection

virtual void connect2in (  int  input, int  hidden )  [virtual]

Connects an input and a hidden neuron with a random weight. Parameters: input  The index of the input neuron hidden  The index of the hidden neuron in the hidden layer

virtual void connect2out (  int  hidden, int  output, real  weight )  [virtual]

Connects a hidden and an output neuron with the given weight. Parameters: hidden  The index of the hidden neuron in the hidden layer output  The index of the output neuron in the output layer weight  The weight of the connection

virtual void connect2out (  int  hidden, int  output )  [virtual]

Connects a hidden and an output neuron with a random weight. Parameters: hidden  The index of the hidden neuron in the hidden layer output  The index of the output neuron in the output layer

virtual void connectAll (   )  [virtual]

Completely connects the network. All inputs are connected to all hidden neurons and all hidden neurons to all output neurons

virtual const char* getClassName (   )  [virtual]

Returns "TwoLayerNetwork".

---

The documentation for this class was generated from the following file:

• TwoLayerNetwork.h

---