neural network and inputs - page 2
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
It's a question of "give me the keys to the flat where the money is". ))))
But let me give you an example:
Everyone well-known Bettinger's Expert Advisor, which won the championship, was trading on Euro-dollar, and the network received information from Euro-yen, because at that time it contained some information about the movement of Euro-dollar and the neural network perfectly highlighted these signals for buying or selling.
But over time, the euro-yen has lost its relevance to the euro-dollar. Then one of well-known indices worked well.
What now - don't know, have to look up..... as bread seekers....))))
О! Leonid! Merry Christmas! Success in business and good health!
How's the project going?
http://www.neuroproject.ru/demo.php?
I wanted to get closer to neuro-experiments myself.
What do you think is relevant at the moment, can you share your ideas on neuro?
Thank you.
Hello!
Merry Christmas to you too! And good luck!
I haven't talked to them in a while..... don't know....
I think it's relevant to look for simple and uncomplicated TCs that are based on simple patterns.
Like, I've shared all sorts of ideas a lot, but only on the forum, it usually doesn't find support, for a variety of reasons.
Hello!
Merry Christmas to you too! And good luck!
It's been a long time since I've spoken to them..... don't know....
I think it's relevant to look for simple and uncomplicated TCs that are based on simple patterns.
Like, I've shared a lot of all sorts of ideas, but only on the forum, it usually doesn't find support, for a variety of reasons.
I see...
More contact later...
On too simple a pattern, there's writing that looks a bit "vulgar" (from the Hunt for Herczyk interview)...
I think it's a bit scary to bet money on simple patterns...
But here is an example:
Bettinger's well-known Expert Advisor that won the Championship, was trading in euro-dollar, and the network got information from the euro-yen, as it contained some information at the time And the network receives information about the movement of euro-dollar and the neural network perfectly separated these signals for buying or selling.
But over time, the euro-yen has lost relevance. for the Euro dollar. Further, one of the indices known to all worked well.
What now - don't know, have to look up..... like bread look up....))))
)))Yes, it's funny.
This effect has long been discovered and described. It's called - "floating" correlation between financial instruments. The correlation coefficient has decreased and the neural network has ceased to produce the correct signals.
I just do not know why a neural network for conventional paired trading should be built....
)))Yes, funny.
This effect was discovered and described a long time ago. It is called "floating" correlation between financial instruments. The correlation coefficient has decreased and the neural network has ceased to produce correct signals.
I just do not understand why a neural network for conventional paired trading should be built....
Your reasoning is wrong because correlation and patterns are absolutely different things.
Instruments can be totally uncorrelated but have patterns.
Your reasoning is flawed because correlation and patterns are completely different things.
Instruments can be completely uncorrelated but have patterns.
Bullshit.
Correlation is a pattern.
"The network was getting information from the euro-yen as it contained some information about the movement of the euro-dollar at the time and the neural network was excellent at picking out those buy or sell signals . " - simple correlation.
"But over time, the euro-yen lost relevance to the euro-dollar. Further, one of the indices known to all worked well." - over time, the correlation decreased and the euro-dollar-index correlation increased.
All this has been known for a long time.
The network is rather predictive, the output has two values in the range [-1;1]. At first I feed 8 values to the input of the network, then I normalize it to the range [-1;1] and don't shift zero. Then I normalize by weights and layers in the same way. The output is a forecast of the two nearest fractals with their relative position relative to the 0 bar. There is no quantitative binding. That is, if the output is -1 and 0.5, it means that the nearest fractal is two times lower than Open[0] and the next fractal is higher than Open[0]. And by analogy, if values are 0.3 and 1 then both fractals are higher than Open[0].
Interesting way of putting it... What is the network? And on the input accordingly some disposition of previous fractals? Just as for me, your two outputs are rather a problem for two networks...
I normalize to range [-1;1] without moving zero. And then by weights and layers I normalize in the same way.
So, the usual normalization on the maximum? I.e. 10; 1; -5 normalized to 1; 0.1; -0.5?
It's not very clear about the normalisation of the weights. Do you normalise them in the same way too? Equally as intermediate layer-by-layer results? Or have I misunderstood something? If correct, then I'm afraid you're going to have some stones here.
And the questions are strange in your opinion. In my head (mind), everything is clear and understandable. And the schemes, and teaching methods, and training, and interpretations. But when it comes to describing the machine - stupor.
The questions are strange out of context. What can you advise about the output without knowing the type of network and its task? The same goes for the input...
Correlation is a pattern.
But not every pattern shows itself in a correlation (even first-year students know this:). It may well be that there is a correlation, but there is zero correlation.
Yeah, so?
What was that about? Why?