Daily Market Analysis - The second day of the forecast is coming to an end - MQL4 and MetaTrader 4

Denis Timoshin 2009.11.24 14:11 #761

mpeugep писал(а) >>

I used to write EAs on it, so I got used to it.

do you want to determine the effectiveness of your forecasting method?

Дима 2009.11.24 14:20 #762

Yes - by testing, which is what I do...

Сергей 2009.11.24 21:17 #763

The second day of the forecast is coming to an end. To recap, the basic "attractors" (i.e. 200 trajectories formed the following "average" trajectories)

Identification unit being tested, instead of the 'ideal trajectory', chose the next one on Saturday, equals (it's not that difficult with them, in terms of choices, usually waiting for the first counts ...):

Throwing the track.

PS: Yay, I'm cool, I've implemented another feature in MT, - "skylining indicator", er I'm not entirely sure it works correctly

double getScaleParametr(double signal[])
{
   int i;
   int n;
   int m;
   
   double h;

   // +------------------------------------------------------+
   // |                       Интегрирование временного ряда |
   // +------------------------------------------------------+
   
   int N;
   int Nsignal;
   
   double delta[];
   
   Nsignal=ArraySize( signal);
   N= Nsignal-1;
   
   ArrayResize( delta, N);
   ArrayInitialize( delta, 0.0);
   
   i=0;
   
   for( n=0; n<= Nsignal-1; n++)
   {
      if( n>0)
      {
         delta[ i]=MathAbs( signal[ n]- signal[ n-1]);
         i= i+1;
      }
   }
   
   // +------------------------------------------------------+
   // |                       Расчет накопленного отклонения |
   // +------------------------------------------------------+
   
   double Mu;
   double SUM;
   double accumulateDeviation[];

   ArrayResize( accumulateDeviation, N);
   ArrayInitialize( accumulateDeviation, 0.0);
   
   Mu= getMu( delta);
   SUM=0.0;
   
   for( n=0; n<= N-1; n++)
   {
      SUM= SUM+( delta[ n]- Mu);
      accumulateDeviation[ n]= SUM;
   }
   
   // +------------------------------------------------------+
   // |                          Итерационный расчет матрицы |
   // +------------------------------------------------------+   
   
   int window;
   int s;
   int k;
   int M;
   int zone;
   
   double a;
   double b;
   
   double x[];
   double y[];
   
   double T[];
   double F[];
   
   double trend[];
   double regress[];
   
   zone=5;

   ArrayResize( T, N- zone);
   ArrayInitialize( T, 0.0);   

   ArrayResize( F, N- zone);
   ArrayInitialize( F, 0.0);   
   
   i=0;
   
   for( window= zone; window<= N; window++)
   {
      s=MathFloor( N/ window);

      ArrayResize( trend, s* window);      
      ArrayInitialize( trend, 0.0);
      
      for( n=0; n<= s-1; n++)
      {
         ArrayResize( x, window);
         ArrayInitialize( x, 0.0);

         ArrayResize( y, window);
         ArrayInitialize( y, 0.0);
         
         for( m=0; m<= window-1; m++)
         {
            x[ m]= m;
            y[ m]= accumulateDeviation[ m+ n* window];
         }

         ArrayResize( regress, 2);
         ArrayInitialize( regress, 0.0);
         
         getLineRegression( regress, x, y);
         
         a= regress[0];
         b= regress[1];

         for( m=0; m<= window-1; m++)
         {
            trend[ m+ n* window]= a+ b* x[ m];
         }
      }

      SUM=0.0;
      
      M=ArraySize( trend);
      
      for( k=0; k<= M-1; k++)
      {
         SUM= SUM+MathPow(( accumulateDeviation[ k]- trend[ k]), 2);
      }
      
      T[ i]=MathLog( window);
      F[ i]=MathLog(MathSqrt((1.0/ M)* SUM));

      i= i+1;
   }

   ArrayResize( regress, 2);
   ArrayInitialize( regress, 0.0);
   
   getLineRegression( regress, T, F);
   h= regress[1];
   
   return( h);
}