private static double GetTotalError(BackPropNeuralNet bnn, bool print)
{
double totalError = 0d, totalErrorOnTestData = 0d;
int score = 0, outof = 0;
for (int i = 0; i < GameData.GameCount; i++)
{
double[] whoPlayed = GameData.GetWhoPlayed(i);
double predictedResult = GameData.RevConv(bnn.ComputeOutputs(whoPlayed)[0]);
double actualResult = GameData.RevConv(GameData.GetGameResult(i));
totalError += (predictedResult - actualResult) * (predictedResult - actualResult);
if (print)
{
if (i == GameData.GameCount - testSize)
Console.WriteLine("---------------------");
Console.WriteLine("Predicted {0:0.00} was {1:0.00} ", predictedResult, actualResult);
if (i >= GameData.GameCount - testSize)
{
totalErrorOnTestData += Math.Abs(predictedResult - actualResult) * Math.Abs(predictedResult - actualResult);
if (Math.Sign(predictedResult) == Math.Sign(actualResult))
score++;
outof++;
}
}
}
if (print)
{
Console.WriteLine("MSE is: {0:0.000}", totalError / (double)GameData.GameCount);
Console.WriteLine("MSE test is: {0:0.000}", totalErrorOnTestData / (double)outof);
Console.WriteLine("Accuracy: {0:0.0}%", (score * 100) / outof);
}
return totalError;
}
static void Main(string[] args)
{
Random rnd = new Random();
GameData.Build();
int numInput = GameData.PlayerCount;
int numHidden = 12; //(int)(GameData.PlayerCount * 0.1d);
int numOutput = 1;
int numWeights = (numInput * numHidden) + (numHidden * numOutput) + (numHidden + numOutput);
Console.WriteLine("Creating a " + numInput + "-input, " + numHidden + "-hidden, " + numOutput + "-output neural network");
Console.WriteLine("Using hard-coded tanh function for hidden layer activation");
Console.WriteLine("Using hard-coded log-sigmoid function for output layer activation");
BackPropNeuralNet bnn = new BackPropNeuralNet(numInput, numHidden, numOutput);
Console.WriteLine("\nGenerating random initial weights and bias values");
double[] initWeights = new double[numWeights];
for (int i = 0; i < initWeights.Length; ++i)
initWeights[i] = (rnd.NextDouble() - 0.5d) * 0.1d;
Console.WriteLine("\nInitial weights and biases are:");
Helpers.ShowVector(initWeights, 3, 8, true);
Console.WriteLine("Loading neural network initial weights and biases into neural network");
bnn.SetWeights(initWeights);
double learnRate = 0.02; // learning rate - controls the maginitude of the increase in the change in weights.
double momentum = 0.0001; // momentum - to discourage oscillation.
Console.WriteLine("Setting learning rate = " + learnRate.ToString("F2") + " and momentum = " + momentum.ToString("F2"));
int maxEpochs = 1000000;
double errorThresh = 0.00001;
Console.WriteLine("\nSetting max epochs = " + maxEpochs + " and error threshold = " + errorThresh.ToString("F6"));
int epoch = 0;
double error = double.MaxValue;
Console.WriteLine("\nBeginning training using back-propagation\n");
while (epoch < maxEpochs) // train
{
int gameNum = rnd.Next(GameData.GameCount - testSize); //leave 10 games for prediction
double[] whoPlayed = GameData.GetWhoPlayed(gameNum);
double result = GameData.GetGameResult(gameNum);
double[] predictedResult = bnn.ComputeOutputs(whoPlayed);
bnn.UpdateWeights(new[] { result }, learnRate, learnRate);
++epoch;
if (epoch % 1000 == 0)
{
error = GetTotalError(bnn, false);
if (error < errorThresh)
{
Console.WriteLine("Found weights and bias values that meet the error criterion at epoch " + epoch);
break;
}
Console.WriteLine("epoch = " + epoch);
Console.WriteLine("error = " + error);
}
} // train loop
double[] finalWeights = bnn.GetWeights();
Console.WriteLine("");
Console.WriteLine("Final neural network weights and bias values are:");
Helpers.ShowVector(finalWeights, 5, 8, true);
GetTotalError(bnn, true);
Console.ReadLine();
//int[] team1 = new int[5];
//int[] team2 = new int[5];
//team1[0] = GameData.GetIndexOfPlayer("Luke");
//team1[1] = GameData.GetIndexOfPlayer("Daniel");
//team1[2] = GameData.GetIndexOfPlayer("MC");
//team1[3] = GameData.GetIndexOfPlayer("Christian");
//team1[4] = GameData.GetIndexOfPlayer("Pete");
//team2[0] = GameData.GetIndexOfPlayer("Ben");
//team2[1] = GameData.GetIndexOfPlayer("Elwood");
//team2[2] = GameData.GetIndexOfPlayer("Nigel");
//team2[3] = GameData.GetIndexOfPlayer("Clive");
//team2[4] = GameData.GetIndexOfPlayer("Harry");
//double[] teamsVector = BuildTeam(team1, team2);
//double[] guess = bnn.ComputeOutputs(teamsVector);
//Console.WriteLine("Guessed result {0:0.00}", GameData.RevConv(guess[0]));
//Console.ReadLine();
}