Hebb neural network

• Download Neural_Network.zip - 51.0 KB

Introduction

In the name of God.

This a sample for OCR(Optical Character Recognition) in c#.

At this Program use the Hebb Algorithm.

Background

Donald Hebb in 1949 write a Article for single-layer problem.

Using the code

int y=0; int j;
int[,] total ={
{1,-1,-1,-1,1,-1,1,-1,1,-1,-1,-1,1,-1,-1,-1,1,-1,1,-1,1,-1,-1,-1,1},
{-1,1,1,1,-1,1,-1,-1,-1,1,1,-1,-1,-1,1,1,-1,-1,-1,1,-1,1,1,1,-1 }
}; 
int[] temp = new int[25];
for (int i = 0; i < 2; i++)
{
for ( j = 0; j < 25; j++)
temp[j] = total[i,j];
switch(i){
case 0: y=1;break;
case 1: y=-1;break;
}
for (j = 0; j < 25; j++)
{
weight[j] = weight[j] + temp[j] * y;
bias = bias + y;
}
}
for (j = 0; j < 25; j++)
{
weights.Items.Add("w"+j.ToString()+" : "+ weight[j].ToString());
}
weights.Items.Add("bias : "+bias.ToString());