Log in

BPNN Predictor JMA smoothing with Cycle period adaptation

BPNN Predictor JMA smoothing with Cycle period adaptation
By JohnLast 3168 days ago Comments (3)

This is a hybrid indicator cpmbining cycle nalysis, jurik smoothing and Neural net prediction with back propagation.

This is for short term prediction. The Cycle period would analyse the cycle period and the jjma would smooth the data before transferring to analyse by the neural net.

In theory it should eliminate some noise compared to the regular BPNN Cycle period neural net.

I think I should add the possibility to modify the niquist factor it is set to 1.5 of dominant cycle.

In this RAR file I add all the necessary files for the installation.

 

Comments

  • jaguar1637 3168 days ago

    Well, I tried to create an Inverse fisher of a PFE indicator, for putting it inside an BPNN Predictor . BUt it failed !! lol

     

     

  • JohnLast 3168 days ago

    Maybe there is something to do with the first line

    #property  indicator_separate_window

    or 

    #property indicator_chart_window 

    As the BPNN is with #property indicator_chart_window

    This is the easiest troubleshooting that comes to my mind.

  • JohnLast 3168 days ago

    I just made a mod with the SSA end - pointed smoothing.