neural_net::Experimental_training_weight

neural_net::Experimental_training_weight< Value_type, Iteration_type, Network_function_type, Space_function_type, Network_topology, Space_topology, Index_type, Parameter_type > Class Template Reference

Template class for the generalize training functions that calculates generalizeg weight in experimental way. More...

#include <generalized_training_weight.hpp>

Inheritance diagram for neural_net::Experimental_training_weight< Value_type, Iteration_type, Network_function_type, Space_function_type, Network_topology, Space_topology, Index_type, Parameter_type >:

[legend]Collaboration diagram for neural_net::Experimental_training_weight< Value_type, Iteration_type, Network_function_type, Space_function_type, Network_topology, Space_topology, Index_type, Parameter_type >:

[legend]

Public Member Functions

template<typename Value_type_2, typename Iteration_type_2, typename Network_function_type_2, typename Space_function_type_2, typename Network_topology_2, typename Space_topology_2, typename Index_type_2, typename Parameter_type_2> Experimental_training_weight (const Experimental_training_weight< Value_type_2, Iteration_type_2, Network_function_type_2, Space_function_type_2, Network_topology_2, Space_topology_2, Index_type_2, Parameter_type_2 > &experimental_training_weight) throw ()
Experimental_training_weight (const Network_function_type &n_f, const Space_function_type &s_f, const Network_topology &n_t, const Space_topology &s_t, const Parameter_type &parameter_0_, const Parameter_type &parameter_1_) throw ()
const Space_function_type::value_type operator() (Value_type &weight, const Value_type &value, const Iteration_type &iteration, const Index_type &c_1, const Index_type &c_2, const Index_type &v_1, const Index_type &v_2) throw ()

Data Fields

Network_function_type network_function
Network_topology network_topology
Parameter_type parameter_0
Parameter_type parameter_1
Space_function_type space_function
Space_topology space_topology

Detailed Description

template<typename Value_type, typename Iteration_type, typename Network_function_type, typename Space_function_type, typename Network_topology, typename Space_topology, typename Index_type, typename Parameter_type>
class neural_net::Experimental_training_weight< Value_type, Iteration_type, Network_function_type, Space_function_type, Network_topology, Space_topology, Index_type, Parameter_type >

Template class for the generalize training functions that calculates generalizeg weight in experimental way.

Parameters:

	Value_type	is a type of values.
	Iteration_type	is a type of interation counter.
	Network_function_type	is a type of function that will return proper value based on network topology.
	Space_funtion_type	is a type of function that will return proper value based on space topology.
	Network_topology	is a type of function that computes distances between neurons based on network topology.
	Space_topology	is a type of function that computes distance between value and weight in proper topology.
	Index_type	is a type of index in the neural network container.
	Parameters_type	is a type of the parameters for experimenntal training. $y= ( p_1 \cdot n_f (n_t (c_1,c_2,v_1,v_2)) - p_0 ) \cdot s_f (s_t (x,w))$

Constructor & Destructor Documentation

template<typename Value_type, typename Iteration_type, typename Network_function_type, typename Space_function_type, typename Network_topology, typename Space_topology, typename Index_type, typename Parameter_type>

template<typename Value_type_2, typename Iteration_type_2, typename Network_function_type_2, typename Space_function_type_2, typename Network_topology_2, typename Space_topology_2, typename Index_type_2, typename Parameter_type_2>

neural_net::Experimental_training_weight< Value_type, Iteration_type, Network_function_type, Space_function_type, Network_topology, Space_topology, Index_type, Parameter_type >::Experimental_training_weight ( const Experimental_training_weight< Value_type_2, Iteration_type_2, Network_function_type_2, Space_function_type_2, Network_topology_2, Space_topology_2, Index_type_2, Parameter_type_2 > & experimental_training_weight ) throw () [inline]

Copy constructor.

Member Function Documentation

template<typename Value_type, typename Iteration_type, typename Network_function_type, typename Space_function_type, typename Network_topology, typename Space_topology, typename Index_type, typename Parameter_type>

const Space_function_type::value_type neural_net::Experimental_training_weight< Value_type, Iteration_type, Network_function_type, Space_function_type, Network_topology, Space_topology, Index_type, Parameter_type >::operator() ( Value_type & weight,

const Value_type & value,

const Iteration_type & iteration,

const Index_type & c_1,

const Index_type & c_2,

const Index_type & v_1,

const Index_type & v_2

) throw () [inline]

Function computes generalized weight for training proces. This weight is not the same as weights in neural network.
Parameters:

weight is a weight from neural network.

value is a input value (value that trains network).

iteration is a number of training steps could be number of training data sample.

c_1 is a row number (position) in the network of the central neuron.

c_2 is a column number (position) in the network of the central neuron.

v_1 is a row number (position) in the network of the trained neuron.

v_2 is a column number (position) in the network of the trained neuron.
$y= ( p_1 \cdot n_f (n_t (c_1,c_2,v_1,v_2)) - p_0 ) \cdot s_f (s_t (x,w))$
where x is value, w is neuron weight, $p_1$ is scaling parameter, $p_0$ is shifting parameter.

Field Documentation

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

generalized_training_weight.hpp

Generated on Wed Jun 28 14:39:54 2006 for Kohonen Neural Network Library Demo by

1.4.6