sigmoid_example.cpp

神经网络vc源代码神经网络的VC++代码呀

/*
    nn-utility (Provides neural networking utilities for c++ programmers)
    Copyright (C) 2003 Panayiotis Thomakos

    This library is free software; you can redistribute it and/or
    modify it under the terms of the GNU Lesser General Public
    License as published by the Free Software Foundation; either
    version 2.1 of the License, or (at your option) any later version.

    This library is distributed in the hope that it will be useful,
    but WITHOUT ANY WARRANTY; without even the implied warranty of
    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
    Lesser General Public License for more details.
*/
//To contact the author send an email to [e-mail]

/*Demonstrates a sigmoid network that learns to classify four letters.*/

#include <nn-utility.h>

using namespace nn_utility;

BITMAP<float> *bit = new BITMAP<float>( "bit.txt" );
typedef nn_utility_functions<float>::VECTOR VECTOR;
 
class sigmoid_example : public nn_utility_functions<float>{
	public:
		typedef float VECTOR[NN_UTIL_SIZE];
		
		void GetInput( int interation, VECTOR &send, VECTOR &target );
		bool CheckTrain( VECTOR output, VECTOR target, int length );
};

VECTOR letters[4];

void sigmoid_example::GetInput( int interation, VECTOR &send, VECTOR &target ){
	CopyVector( send, letters[interation % 4], 25 );

	switch ( interation % 4 ){
		case 0 : LoadVectorf( target, 9, 0.9, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1 ); break;
		case 1 : LoadVectorf( target, 9, 0.1, 0.9, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1 ); break;
		case 2 : LoadVectorf( target, 9, 0.1, 0.1, 0.9, 0.1, 0.1, 0.1, 0.1, 0.1, 0.1 ); break;
		case 3 : LoadVectorf( target, 9, 0.1, 0.1, 0.1, 0.9, 0.1, 0.1, 0.1, 0.1, 0.1 ); break;
	}
}

bool sigmoid_example::CheckTrain( VECTOR output, VECTOR target, int length ){
	for ( int ir = 0; ir < length; ir++ ){
		if ( output[ir] < target[ir]-0.2 || output[ir] > target[ir]+0.2 ){
			return true; }
	}
	return false;
}

void SetRandomWeights( layer<float> **in, int row, int col ){
	for ( int i = 0; i < col; i++ ){
		float it2 = rand() % 10;
		(*in)->weight[i] = it2/10*( (rand() % 2) == 0 ? -1 : 1 );
		for ( int e = 0; e < row; e++ ){
			float it = rand() % 10;
			(*in)->matrix[e][i] = it/10*( (rand() % 2) == 0 ? -1 : 1 );
		}
	}
}

int Largest( VECTOR output, int length ){
	int index = 0;
	for ( int i = 1; i < length; i++ ){
		if ( output[i] > output[index] ){ index = i; }
	}
	return index;
}

void classify_result( VECTOR input, VECTOR output ){
	bit->Print( input, 5, 25 );
	cout << "Network fired: ";
	int big = Largest( output, 9 );
	switch( big ){
		case 0 : cout << "A\n"; break;
		case 1 : cout << "B\n"; break;
		case 2 : cout << "C\n"; break;
		case 3 : cout << "D\n"; break;
		default: cout << "unknown\n"; break;
	};
}

int main(){
	sigmoid_example derived;

	SIGMOID *hidden = new SIGMOID();
	hidden->define( 25,6 );
	SIGMOID *classify = new SIGMOID();
	classify->define( 6,9 );

	layer<float> *ppHidden   = hidden;
	layer<float> *ppClassify = classify;
	
	SetRandomWeights( &ppHidden, 25,6 );
	SetRandomWeights( &ppClassify, 6,9 );
	
	for ( int i = 0; i < 4; i++ ){
		bit->readbitmap( letters[i], i );
		hidden->sigmoid( letters[i], 25 ); }
	
	derived.Insert( &ppHidden, &ppClassify );

	VECTOR FINAL;
	derived.ClearVector( FINAL, 9 );
	
	derived.train( &ppHidden, &ppClassify, 100, 0.9F, true );
	cout << "\n\n";
	cout << "Trained for 100 epochs at a learning rate of 0.3 under supervised learning.\n";
	cout << "The bitmap noise() function added noise to the following letters and the network\n";
	cout << "classified them according to the results below each 5x5 bitmap\n";

	cout << '\n';
	for ( int go = 0; go < 4; go++ ){
		cout << "Given noisy input pattern:\n";
		bit->noise( letters[go], 25, 0, 10 );
		hidden->FeedForward( letters[go], FINAL );
		classify_result( letters[go], FINAL );
	}

	return 0;
}