Hello Mr. Mohamed. I suppose, M15 is more optimal time-frame than M5. M5 is very noisy interval. 1-3 nearest points are more reliable data.

In general the next param sets on start:

1. SSA Fast Trend Forecast 2.5:
   Algorithm: Recurrent forecast,
   N: Data fragment = 256,  
   Time-dependent lag = N/3, 
   Trend    high-freq. limit= 0.25,
   Forecast high-freq. limit= 0.25,
   Forecast transform = S[i]/Max(:),
   Forecast smoothing = Smoothing MA(3).
2. SSACD Forecast (Limited)  2.5:
   Algorithm: Recurrent forecast,
   N: Data fragment = 512,  
   Time-dependent lag = N/4, 
   FastTrend high-freq. limit = 0.4
   SlowTrend high-freq. limit= 0.6
   Signal SMA period = 4
   Data preparation = {ln(S[i]-Smin+1)}/Max(:)
   Forecast preparation = S[i] /Max(:)
   Forecast smoothing = Smoothing MA(3).
3. SSA Stochastic (Limited) 2.0:
   Algorithm: Recurrent forecast,
   N: Data fragment = 256,  
   Time-dependent lag = N/4, 
   %K high-freq. limit = 0.3,
   %D high-freq. limit = 0.6,
   Data preparation = S[i] /Max(:),
   Forecast smoothing = Smoothing MA(3).