What is the perspective in your opinion?

 

I would like to hear (preferably from people who have drawn conclusions from their own experience) and maybe discuss promising methods for analysing financial markets (neural networks or cluster indicators for example - and why).

 

The main thing is to look for informative signs, otherwise...

because a neural network will synthesise anything you give it, that's what it's all about and

cluster analysis will cluster anything you give it, that's what it's all about.

That is the main thing.

 
and then finding out the measure of usefulness of the results of your advising system
 
IMHO, there are no indicators in the BEAST of analysis methods, i.e. statistics, probabilities, etc.
 

Yeah) lots of tips, no one wants to show off?)

I'd like some complex formulas and twists =)

 
learn the TERVER
 
Summer:

I would like to hear (preferably from people who have drawn conclusions from their own experience) and maybe discuss promising methods for analysing financial markets (neural networks or cluster indicators for example - and why).



I am a Forex trader not so long ago. Started studying MQL at the beginning of March 2009. Wrote a bunch of indicators. After a year I realized that it was not the right one. I started to use Expert Advisors in NS. I am using one on real account as of today. I am already working on my forex trading. You may watch the last week on demo.
Files:
statement.zip  7 kb
 
icas:

I am in forex recently. I`ve started to study MQL since the beginning of March 2009. I`ve written a lot of indicators. After a year I realized - this is not the right one. Started to use Expert Advisors on NS. I am using one on real account as of today. I am already working on my forex trading. You may watch the last week on demo.



I have a couple of friends who know me. I have a couple of friends in the NS - they do such wonders with profit.

But the topic is not simple and it is necessary to prepare the NS correctly to get something, I tried to use the library FANN - on the history does a fit, but beyond the fit - then ... the curve is going in the wrong direction =)

 
Summer:

Yeah) lots of tips, no one wants to show off?)

I'd like some complex formulas and twists =)

Full probability formula =)

 
Summer:



NS is yes. I have a couple of friends who work in NS - they do such wonders with profit.

But the topic is not simple and it is necessary to prepare the NS correctly to get something, I tried the library FANN - on the history it makes a fit, but beyond the fit... the curve is going in the wrong direction =)


I do not use other people's codes, because there is no way to understand other people's thoughts (I would like to understand my own). I've read on a forum about NS and wrote a monster (48 optimization steps, 12-14 parameters in each, 31 hours of processor time).
 
Those who know the formula do not need NS.
Reason: