static double[][] MakeAllData(int numInput2, int numHidden3, int numHidden4, int numOutput2, int numRows2, int seed2)
{
Random rnd = new Random(seed2);
int numWeights = (numInput2 * numHidden3) + numHidden4 +
(numHidden4 * numOutput2) + numOutput2;
double[] weights = new double[numWeights]; // actually weights & biases
for (int i = 0; i < numWeights; ++i)
{
weights[i] = 20.0 * rnd.NextDouble() - 10.0; // [-10.0 to 10.0]
}
Console.WriteLine("Generating weights and biases:");
ShowVector(weights, 2, 10, true);
double[][] result = new double[numRows2][]; // allocate return-result
for (int i = 0; i < numRows2; ++i)
{
result[i] = new double[numInput2 + numOutput2]; // 1-of-N in last column
}
NeuralNetwork1 gnn =
new NeuralNetwork1(numInput2, numHidden3, numHidden4, numOutput2, numRows2, seed2); // generating NN
gnn.SetWeights(weights);
for (int r = 0; r < numRows2; ++r) // for each row
{
// generate random inputs
double[] inputs = new double[numInput2];
for (int i = 0; i < numInput2; ++i)
{
inputs[i] = 20.0 * rnd.NextDouble() - 10.0; // [-10.0 to -10.0]
}
// compute outputs
double[] outputs = gnn.ComputeOutputs(inputs);
// translate outputs to 1-of-N
double[] oneOfN = new double[numOutput2]; // all 0.0
int maxIndex = 0;
double maxValue = outputs[0];
for (int i = 0; i < numOutput2; ++i)
{
if (outputs[i] > maxValue)
{
maxIndex = i;
maxValue = outputs[i];
}
}
oneOfN[maxIndex] = 1.0;
// place inputs and 1-of-N output values into curr row
int c = 0; // column into result[][]
for (int i = 0; i < numInput2; ++i) // inputs
{
result[r][c++] = inputs[i];
}
for (int i = 0; i < numOutput2; ++i) // outputs
{
result[r][c++] = oneOfN[i];
}
} // each row
return(result);
} // MakeAllData
static void Main(string[] args)
{
Console.WriteLine("\nbegin ANN.\n");
int numInput2 = 4; // number features
int numHidden3 = 5;
int numHidden4 = 5;
int numOutput2 = 3; // number of classes for Y
int numRows2 = 1000;
int seed2 = 1; // gives nice demo
Console.WriteLine("\nGenerating " + numRows2 +
" artificial data items with " + numInput2 + " features");
double[][] allData = MakeAllData(numInput2, numHidden3, numHidden4, numOutput2,
numRows2, seed2);
Console.WriteLine("Done");
//ShowMatrix(allData, allData.Length, 2, true);
Console.WriteLine("\nCreating train (80%) and test (20%) matrices");
double[][] trainData2;
double[][] testData2;
SplitTrainTest(allData, 0.80, seed2, out trainData2, out testData2);
Console.WriteLine("Done\n");
Console.WriteLine("Training data:");
ShowMatrix(trainData2, 4, 2, true);
Console.WriteLine("Test data:");
ShowMatrix(testData2, 4, 2, true);
Console.WriteLine("Creating a " + numInput2 + "-" + numHidden3 +
"-" + numHidden4 + "-" + numOutput2 + " neural network");
NeuralNetwork1 nn = new NeuralNetwork1(numInput2, numHidden3, numHidden4, numOutput2, numRows2, seed2);
int maxEpochs = 1000;
double learnRate = 0.05;
double momentum = 0.01;
Console.WriteLine("\nSetting maxEpochs = " + maxEpochs);
Console.WriteLine("Setting learnRate = " + learnRate.ToString("F2"));
Console.WriteLine("Setting momentum = " + momentum.ToString("F2"));
Console.WriteLine("\nStarting training");
double[] weights = nn.Train(trainData2, maxEpochs, learnRate, momentum);
Console.WriteLine("Done");
Console.WriteLine("\nFinal neural network model weights and biases:\n");
// ShowVector(weights, 2, 10, true);
//double[] y = nn.ComputeOutputs(new double[] { 1.0, 2.0, 3.0, 4.0 });
//ShowVector(y, 3, 3, true);
double trainAcc = nn.Accuracy(trainData2);
Console.WriteLine("\nFinal accuracy on training data = " +
trainAcc.ToString("F4"));
double testAcc = nn.Accuracy(testData2);
Console.WriteLine("Final accuracy on test data = " +
testAcc.ToString("F4"));
Console.WriteLine("\nBegin van Vector.\n");
double[][] trainData = new double[4][];
trainData[0] = new double[] { 2.0, 2.0 }; // data cannot be perfectly classified
trainData[1] = new double[] { 3.5, -4.0 };
trainData[2] = new double[] { 4.0, -5.5 };
trainData[3] = new double[] { 4.5, -2.0 };
int[] Y = new int[4] {
0, 1, 1, 1
};
Console.WriteLine("\nTraining data: \n");
ShowTrainData(trainData, Y);
double[] weights2 = null;
double bias = 0.0;
double alpha = 0.001;
int maxEpochs2 = 500;
Console.Write("\nSetting learning rate to " + alpha.ToString("F3"));
Console.WriteLine(" and maxEpochs to " + maxEpochs2);
Console.WriteLine("\nBeginning training the perceptron");
Train(trainData, alpha, maxEpochs2, Y, out weights2, out bias);
Console.WriteLine("Training complete");
Console.WriteLine("\nBest percetron weights found: ");
ShowVector(weights2, 4);
Console.WriteLine("\nBest perceptron bias found = " + bias.ToString("F4"));
double acc = Accuracy(trainData, weights2, bias, Y);
Console.Write("\nAccuracy of the perceptron on the training data = ");
Console.WriteLine(acc.ToString("F2"));
Console.WriteLine("\nDone\n");
Console.ReadLine();
} //Main
