public void TrainSLP()
{
perceptronModel.Fit(0.01f, int.Parse(slpEpoch.text), dataTrain);
slpAcc.text = perceptronModel.Accuracy(dataTrain).ToString();
}
public void TrainMLP()
{
backpropModel.Fit(0.01f, int.Parse(mlpEpoch.text), dataTrain);
mlpAcc.text = backpropModel.Accuracy(dataTrain).ToString();
}
static void Main(string[] args)
{
Console.WriteLine("****************** BIAI PROJECT - Neural Network ******************");
Console.WriteLine("****************** Adam Pałka & Jakub Walczak *******************\n");
//Console.WriteLine(" 0 0 1 = Iris Setosa");
//Console.WriteLine(" 0 1 0 = Iris Versicolor");
//Console.WriteLine(" 1 0 0 = Iris Virginica\n");
Console.WriteLine("INPUT DATA: sepal length, sepal width, petal length, petal width");
Console.WriteLine("INPUT DATA example");
Console.WriteLine("5.1, 3.5, 1.4, 0.2 (Iris Setosa)");
Console.WriteLine("7.0, 3.2, 4.7, 1.4 (Iris Versicolor)");
Console.WriteLine("6.8, 3.0, 5.5, 2.1 (Iris Virginica)");
Console.WriteLine("*******************************************************************\n");
double[][] data = null;
double[][] dataForTest = null;
double[][] dataForTraining = null;
Tools.FileReader(ref data, @"..\..\DATASET.txt"); //reading data from file
DataPrepare.MakeTrainTest(data, ref dataForTest, ref dataForTraining);
DataPrepare.Normalize(dataForTest, new int[] { 0, 1, 2, 3 });
DataPrepare.Normalize(dataForTraining, new int[] { 0, 1, 2, 3 });
Console.WriteLine("Creating neural network...");
const int input = 4;
const int hidden = 7;
const int output = 3;
NeuralNetwork neuralNetwork = new NeuralNetwork(input, hidden, output);
neuralNetwork.InitializeWeights(); //Initializing weights and bias to small random values
//TRAINING
//Settings for training
int epochos = 2000;
double learnRate = 0.05;
double momentum = 0.01;
double weightDecay = 0.0001;
Console.WriteLine("Start training");
DateTime startTime = DateTime.Now;
neuralNetwork.Training(dataForTraining, epochos, learnRate, momentum, weightDecay);
DateTime stopTime = DateTime.Now;
TimeSpan timeDifference = stopTime - startTime;
Console.WriteLine("Training time:" + timeDifference.TotalSeconds);
Console.WriteLine("Training completed!");
double[] weights = neuralNetwork.GetWeights();
//Accuracy
double accurancyTrain = neuralNetwork.Accuracy(dataForTraining);
Console.WriteLine("--------------------Accurancy on Neural Network-------------------- ");
Console.WriteLine("\nAccuracy on training data = " + accurancyTrain.ToString("F2"));
double accurancyTest = neuralNetwork.Accuracy(dataForTest);
Console.WriteLine("Accuracy on test data = " + accurancyTest.ToString("F2"));
//TESTING BY USER
double[] numbers = new double[4];
string userInput;
Console.WriteLine("\n------------------------------------------------------------------- ");
Console.WriteLine("Would you like to test ? [Y/N]");
userInput = Console.ReadLine();
Tools.CheckInputAnswer(userInput);
while (userInput.ToUpper() == "Y")
{
Console.WriteLine("Please enter 4 values [sepal length, sepal width, petal length, petal width]: ");
for (int a = 0; a < 4; a++)
{
try
{
numbers[a] = Convert.ToDouble(Console.ReadLine(), System.Globalization.CultureInfo.InvariantCulture);
}
catch (Exception e)
{
Console.WriteLine("Wrong value, try again!");
}
if (numbers[a] <= 0)
{
Console.WriteLine("Wrong value, try again!");
a--;
}
}
Console.WriteLine(Tools.OutputType(ref neuralNetwork, DataPrepare.NormalizeInput(numbers, data)));
Console.WriteLine("Would you like to test again ? [Y/N]");
userInput = Console.ReadLine();
Tools.CheckInputAnswer(userInput);
}
}
