Dependency statistics in quotes (information theory, correlation and other feature selection methods) - page 56
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
VNG: Есть. Два шикарных стейта. Один из них - подъем с 50 баксов до 10000 за 10 сделок в течение 3-х месяцев. Ни одной убыточной.
10000/50 = 200 = x^10.
x = 1.7.
Bloody hell!
TAdv - Adverse Tactics, presented by multipoints. Screens were posted here by one of the authors.
Vadimcha's channels and swings are models published online by a user called Vadimcha. I won't give any links, google the nickname, you'll find them. I searched for them myself three years ago.
Screenshots with models I posted just above.
My interest is in formalization of these patterns for automation.
This is what I would like to formalize. The screenshot shows the sequence of two black pulses. The task is to calculate the length of the third red pulse and the moment of arrival to the endpoint using TI methods.
If there is a methodology for extracting patterns, then describe step by step what and how to calculate in your terms of reference, hire a programmer at the appropriate level and you will be happy.
Yusuf will not lie.
They are not just dependent, they are extremely dependent! The econometric platitude about Pearson autocorrelation on the first few bars has long been known to me. But it is of no use to me.
Actually, it's pretty much the same thing I did. Only in a different language.
If someone is confused by the language of TI - ok, you can use the language of statistics. Hee-squared, for fuck's sake!
The interpretation is there. Read Wiki:
Our research says otherwise: financial markets are informationally inefficient!Stop, stop, stop....
The question is specific. The article cites an information dependency. A marketplace is not a company of soldiers in height. It is information. There is ACF and there is another formula from TI. Where is the basis that it is better (worse) than ACF? So I too, pick my nose and write a formula in opposition to ACF. ACF reveals trends, which we see on the quotient. And it reveals trends, cycles that are not very visible to the eye. This is an auxiliary tool for further fixation of movements in the analytical form. And what does TI reveal? These alleged TI dependencies affect the movement of the quotient?
10000/50 = 200 = x^10.
x = 1.7.
Bloody hell!
That's a fact.
10000/50 = 200 = x^10.
x = 1.7.
Bloody hell!
You and I have already said that ACF only reveals linear dependencies. The ACF of a quotient no later than the 10th step is destroyed into a statistically indistinguishable figure from zero.
And this formula from TI (or, similarly, chi-square criterion) reveals any dependences, any degree of non-linearity. And the dependencies are not up to the 10th bar, but up to a number in the thousands. Feel the difference?
If there is a methodology for extracting patterns, then describe step by step what and how to calculate in your terms of reference, hire a programmer at the appropriate level and you will be happy.
Yusuf will not lie.
I have no use for hiring a programmer. I do a little sculpting myself and have friends who have never turned me down. The problem is something else - it is not clear how to formalize it. You can see with your head and eyes, but you can't formalize it. But that is not what I mean at the moment. I want to solve exactly the problem of forecasting.
On your own skin.
one hide is not enough for statistics))
We've been hanging out here for nothing. We've been discussing serious and interesting things.....
And this is the limit of the pursuit of perfection, which it isn't.
You and I have already said that ACF only reveals linear relationships. ACF kotir no later than the 10th step is annihilated into a statistically indistinguishable from zero figure.
And this formula from TI (or, similarly, chi-square criterion) reveals any dependences, any degree of non-linearity. And the dependencies are not up to the 10th bar, but up to several thousands. Feel the difference?
I don't get it. ACF draws as much as it puts in, what's the use?
Why can you trust the TI on a large number of bars?