Dependency statistics in quotes (information theory, correlation and other feature selection methods) - page 66
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
You see, in this case, you have to try. And for us to observe.
And your goal would be to show that it's not bullshit.
Vladimir, observe? free shows, in the forum header. click on the dream.
I concede that your goals, to find or make sure of something that isn't bullshit. those are good goals, in my opinion.
My goal and being here, are far less stressful. )))
Vladimir, watch? free shows, in the forum header. click on the dream.
I concede that your goals, to find or make sure of something that isn't dumb. those are good goals, in my opinion.
My goal and being here, are far less stressful. )))
You are not alone. No holder of everlasting principles has ever been able to do anything.
And no wonder. Casus vulgaris.)
By the way, Alexey, you haven't commented on my posted screenshots. About the justification of fractality of the market. Or do you think the returns are taken incorrectly? There is a clear exponential distribution. I can't find a better justification for fractal invariance.
Nikolai, are you talking about the post with the getch quote (who he is on five now, you should know)?
Well... I got hung up on that post and was going to reply later. Good stuff, interesting, must have a closer look.
By non-fractalism in this case I mean violation of any invariance when changing the scale. In the case of the subject of this branch there is no invariance with dependences: there are too many dependences on hours, much less on H4 and very few on days. And this despite exponential distribution of the returns themselves.
How can I explain it... A study of the getch is a construction of a one-dimensional distribution of ZZ knees. It seems to be exponential too, doesn't it?
The graph with the distribution of cities by size is also a univariate distribution.
In both cases, no dependencies between data series seem to have been investigated. To investigate them, you have to handle the joint distributions of the two quantities.
So it seems to me that the univariate distribution of the data alone is not enough to confidently infer the cognitive or physical nature of the phenomenon. This is only an indication of, but not a substantiation of.
I am certainly not claiming that the market is predominantly physical in nature. It just turns out to be more complex than we think. Well the author of the article has those words too:
The lack of a characteristic scale in the parameters of a phenomenon is a signature of the cognitive order governing that phenomenon. It often looks like an apparent fractality in the structure of the phenomenon, but not necessarily so. For example, cities and their populations do not seem to form any apparent fractal structure, but the population of cities does not have a characteristic scale either.
2 Vadimcha:
Perhaps then excessive sympathy to the tester, will gradually disappear, but the understanding of non-randomness of events, will only be restored, in one way or another.
Vadim, thank you, I understand now. Full immersion in the task, so to speak... extremism in action.
This means that the reliability of the prediction is falling. With each step, it gets smaller and smaller. Is that right?
Of course it is.
No, it depends on the model. Take the most famous model in econometrics - cointegration. This model is essentially a model for spread trading, statistical arbitrage and some others. There the error does not accumulate over time. More precisely, it is based on a mechanism that seeks to minimise the accumulated error - the trend is just determined by the accumulated error. The mechanism is called ECM (Error Correction Model), or error correction method.
You are not alone. No one with everlasting principles has ever been able to do anything.
And no wonder. Casus vulgaris.)
Well you're not alone either, Vladimir. )))) the forums get bloated - the brains get dry. (remember the proverb - the progenitor?)
And without "eternally valid principles", what arguments have you used to convince yourself that you can't fish, just because the rod is unformalizable through the selection of a junior period, on smoothing? ;)
Well, you're not alone either, Vladimir. )))) the forums get bloated - the brains get dry... (remember the proverb - the progenitor?)
And without "everlasting principles", what arguments have you used to convince yourself that it's impossible to fish, just because the rod is unformalizable through the selection of a junior period, on smoothing? ;)
Vadim, why did you decide not to fish but to brag about owning a cool non-formalizable rod?
There are fish. Not in this pond, but plenty.
Vadim, why have you decided not to fish, but to brag about owning a cool unformalized rod?
Fish that is.
I've already heard the phrase about formalization in this thread. and there are fish, i agree, and i can argue - i'm not bragging.
i asked a question about the bogeyman. i got some answers, and not all of them were understood.
Thanks nevertheless, the answers were clever and evasive.
Nikolai, are you talking about the post with the getch quote (who he is on five now, you should know)?
Well... I got hung up on that post and was going to reply later. Good stuff, interesting, have to take a closer look.
By non-fractalism in this case I mean the violation of any invariance when changing the scale. In the case of the subject of this branch there is no invariance with dependencies: there are too many dependencies on hours, much less on H4 and very few on days. And this despite the exponential distribution of returns.
How can I make it clear... The getch study is the construction of a one dimensional distribution of the ZZ knees. It seems to be exponential too, doesn't it?
The graph with the distribution of cities by size is also a univariate distribution.
In both cases, no dependencies between data series seem to have been investigated. To investigate them, you have to handle the joint distributions of the two quantities.
So it seems to me that the univariate distribution of the data alone is not enough to confidently infer the cognitive or physical nature of the phenomenon. This is only an indication of, but not a substantiation of.
I am certainly not claiming that the market is predominantly physical in nature. It just turns out to be more complex than we think. Well the author of the article has those words too:
2 Vadimcha:
Vadim, thanks, I understand now. So to speak, a full immersion into the task... extremism in action.
Alexei, thank you for your tolerance and quick answers.
Whogetsch on the five I don't know. However, I can see that he stumbled upon a very interesting result when researching and did not see it.
I realise that my posts in this thread are on the verge of a foul, almost an off-topic. This thread was created for another specific purpose and my attacks cause, how shall I put it mildly... some irritation and resentment. At the same time to create a separate branch too much responsibility, not yet ready.
If I may, a few more questions.
- What is any invariance when the scale changes, sorry, I do not understand. I understand invariance as the presence of a scale factor (in general it can be any number or function), when multiplied by which we obtain a new pattern on a different scale. That is, the affine transformation, which is the manifestation of structuredness in a chaotic flow of data. The problem then comes down to finding such a coefficient. When a pattern is found it is simply multiplied by this coefficient. And such transformation works both "upwards" and "downwards". And that is all.
- Why does it seem? The screenshot shows a clear graph of the exponent.
- If you investigate the relationship between the two quantities.
- why is it so, what is the reason for this statement
- Why two and not three or five or thirty?
- which two quantities
- the joint distribution of two quantities represents a surface. What, are we moving to another raelty?
I was about to end my presence here. But let's wait until tomorrow. Let faa1947 confirm or refute your point with fresh air. My question: econometrics does not postulate axioms, theorems and hypotheses?
Don't bother with nonsense. I am not an econometrician. Econometrics is the application of matstatistics to economics. Matstatistics is widely used in medicine, which today can't be thought of without matstatistics. But there is no appropriate name for it.
In econometrics, I am interested in tools that can fill my pocket. Exactly like some kind of indicator.