// main program, which calls whatever is needed for the analysis
static void Main(string[] args)
{
DateTime GlobalClock = DateTime.Now;
Console.WriteLine("I N I Z I O _ P R O G R A M M A.\n\n\n");
// a dictionary is used to match input files names and integers that are required to the user
Dictionary<int, string> dict = new Dictionary<int, string>
{
{ 0, "WineDS" },
{ 1, "BioMechDS" },
{ 2, "IrisDS" }
};
// being InOutStream static, it need no instantiation
InOutStreams.GoToFolder();
Console.WriteLine(InOutStreams.GetPathOut());
Console.WriteLine("Scelta del DataSet:\nWine DataSet: premere 0\nBioMechanical DataSet: premere 1\nIris DataSet: premere 2");
int FlagDataSet = Convert.ToInt32(Console.ReadLine());
InOutStreams.DirectoryManager(FlagDataSet, dict);
// is worth note that even if DataObject is the heart of the data modelling, prior instantiation
// of the class containing the methds for import is necessary, then the DataObject is instantiated
// and filled in with the data read by the DataSetImport methods
DataSetImport DSImp = new DataSetImport();
DataObject DataSet = DSImp.FeaturesImport();
Console.WriteLine("\n\nPreprocessare i dati? (PCA):\nNo: premere 0\nSi: premere 1");
int FlagProcessing = Convert.ToInt32(Console.ReadLine());
if (FlagProcessing == 1)
{
DataProcessing process = new DataProcessing();
DataSet = process.DoThePCA(DataSet);
DataProcessing.SetFlagProcessing(FlagProcessing);
}
// NOTE: in DataWorkTable.DataProcessing two static field are located. Here, according to the
// user choice to preprocess or not, the if () statement above is considered (or not), BUT to
// the information about data preprocessing choice must be in any case remembered, thus, in the
// if() body the DataProcessing class is instantiated, and eventually the static field about choice
// and number of PCs stored, but if the if() is not executed, such choices are not even present,
// and they are needed in the following. So the static field initialised to 0 is anyway present
// and solves this problem
Console.WriteLine("\n\nPartizione del DataSet?");
Console.WriteLine("TrainingSet = 90% del DataSet e TestSet = 10% del DataSet: premere 0");
Console.WriteLine("TrainingSet = 66% del DataSet e TestSet = 33% del DataSet: premere 1");
int FlagPartition = Convert.ToInt32(Console.ReadLine());
InOutStreams.DirectoryCreation(FlagPartition, DataProcessing.GetFlagProcessing(), DataProcessing.GetPrincipalComponentsNumber());
Console.WriteLine("\n\nDrop folder: \n" + InOutStreams.GetPathOut());
Console.WriteLine("\n\nSingle run o ciclo? (inserire 1 per single run, altrimenti numero di iterazioni)");
int iter = Convert.ToInt32(Console.ReadLine());
for (int i = 0; i < iter; i++)
{
if (iter > 1)
{
Console.WriteLine("Iterazione {0} / {1}", i + 1, iter);
}
DataPartition Partition = new DataPartition(DataSet, FlagPartition);
DataObject[] SetsData = Partition.TestSetExtraction(DataSet);
Partition.PrintPartition(SetsData);
Console.WriteLine("\n\n");
LearningAlgorithmsClass la = new LearningAlgorithmsClass(SetsData, FlagPartition, DataProcessing.GetFlagProcessing(), DataProcessing.GetPrincipalComponentsNumber());
// given all the previous choices, the whole learning machinery is deployed
la.SVMGaussianKernel();
Console.WriteLine("\n\n");
la.SVMPolinomialKernel();
Console.WriteLine("\n\n");
la.TreeLearning();
Console.WriteLine("\n\n");
}
Console.WriteLine("\n\n");
Console.WriteLine("\n\n\nTempo di esecuzione: " + (DateTime.Now - GlobalClock).TotalSeconds + " secondi.");
Console.WriteLine("\nF I N E _ P R O G R A M M A.");
Console.WriteLine("Premi qualsiasi tasto per uscire.");
Console.ReadKey();
}