/// <summary>
/// Converts a List of configurations with its measured value (we take the non-functional property of the global state) into a a learning matrix.
/// </summary>
/// <param name="measurements">The configurations containing the measured values.</param>
public void setLearningSet(List <Configuration> measurements)
{
double[] temparryLearn = new double[measurements.Count];; //measured values
for (int i = 0; i < measurements.Count; i++)
{
this.learningSet.Add(measurements[i]);
double val = 0;
try
{
val = measurements[i].GetNFPValue(GlobalState.currentNFP);
} catch (ArgumentException argEx) {
GlobalState.logError.log(argEx.Message);
val = measurements[i].GetNFPValue();
}
temparryLearn[i] = val;
}
Y_learning = temparryLearn;
Y_learning = Y_learning.T;
//Now, we genrate for each inidividual option the data column. We also remove options from the initial feature set that occur in all or no variants of the learning set
List <Feature> featuresToRemove = new List <Feature>();
foreach (Feature f in this.initialFeatures)
{
if (f.participatingNumOptions.Count > 0)
{
continue;
}
List <Feature> temp = new List <Feature>();
temp.Add(f);
ILArray <double> column = createDataMatrix(temp);
int nbSelections = 0;
int nbDeselections = 0;
for (int i = 0; i < column.Length; i++)
{
if (column[i] == 1)
{
nbSelections++;
}
if (column[i] == 0)
{
nbDeselections++;
}
if (nbSelections > 0 && nbDeselections > 0)
{
break;
}
}
if (nbSelections == this.learningSet.Count)
{
featuresToRemove.Add(f);
}
if (nbDeselections == this.learningSet.Count)
{
featuresToRemove.Add(f);
}
}
foreach (var f in featuresToRemove)
{
this.initialFeatures.Remove(f);
}
}
protected void readlog()
{
Dictionary<string, LinkedList<double>> d = log.Run();
List<string> list = new List<string>(d.Keys);
//O1 = new ILArray<double>[0];
if (d.Keys.Count > 0)
{
foreach (string l in d.Keys)
{
O1 = new ILArray<double>[d[l].Count];
int i = 0;
foreach (double db in d[l])
{
O1[i] = db;
i++;
}
//Task.Factory.StartNew(() => System.Console.WriteLine(ILNumerics.ILMath.fft(O1)));
ILNumerics.ILRetArray<ILNumerics.complex> o;
o = ILNumerics.ILMath.fft(O1);
LinkedList<double> ds = extract_freq(o);
double max = get_max(ds);
double avg = average(ds);
Console.WriteLine("LIST: ");
//print_list(ds);
Console.WriteLine("max: " + max + " avg: " + avg);
if (max - avg > 13000)
{
Task task = new Task(new Action(() => level.ship.Accerlerate()));
task.Start();
task.Wait(2000);
}
//Console.Write("list is: ");
//Task.Factory.StartNew(() => print_list(ds));
}
}
}
private void OnTimedEventData(object source, ElapsedEventArgs e)
{
ILNumerics.ILRetArray<ILNumerics.complex> o = ILNumerics.ILMath.fft(O1);
if (O1.IsEmpty)
{
return;
}
LinkedList<double> ds = extract_freq(o);
//print_list(ds);
Task.Factory.StartNew(() => plot(ds));
int move = scan_list(ds);
switch(move){
case 1:
level.ship.Accerlerate();
break;
case 2:
level.ShipShoot();
break;
case 3:
level.ship.RotateShipLeft(90);
break;
case 4:
level.ship.RotateShipRight(90);
break;
default:
break;
}
O1 = new ILArray<double>[2000];
O1_count = 0;
}
protected override void Initialize()
{
freeCam = false;
currentCamera = 1;
lev = 0;
graphics.PreferredBackBufferWidth = 800;
graphics.PreferredBackBufferHeight = 480;
up_color = Color.Black;
down_color = Color.Black;
left_color = Color.Black;
right_color = Color.Black;
log = new EEG_Logger();
Task.Factory.StartNew(() => start_timers());
O1 = new ILArray<double>[2001];
O1_count = 0;
base.Initialize();
nocounts = 0;
}
