The market is a controlled dynamic system. - page 362
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
Cooper J., McGilliam K. Probabilistic methods for the analysis of signals and systems
pp. 180
We are interested in non-stationary processes. (I underlined)
(With stationary ones everything is clear)Thank you.
It turns out that tau for non-stationary processes can and should be a variable...
It turns out that I'm right, working with Erlang flows... Only in them the ACF will be calculated correctly.
Thank you.
It turns out that tau for unsteady processes can and should be a variable...
It turns out that I'm right, working with Erlang flows... Only in these will the ACF be calculated correctly.
tau is in any case a variable -- both stationary and non-stationary
Erlangs have nothing to do with it.
The tau is the interval between the measured values. If the interval changes, the tau changes.
tau is in any case a variable - both for stationary and non-stationary
erlangs have nothing to do with it.
The tau is the interval between the measured values. The interval changes - the tau changes.
I understood that at different tau, in case of non-stationary process, we get different values of ACF and to set, for example, tau=1 min. and to think, that there is some true ACF in front of me would be wrong. Right?
So, I think when tau varies exponentially, as in Erlang flows, then already these ACF values can be trusted...
Sorry, maybe I'm being silly, but I'm in desperate need of a parameter that calculates process "memory". Really hoping for an ACF...
1) I understand that with different tau, in case of non-stationary process, we get different values of ACF and to set, for example, tau=1 min. and think that I have some true ACF in front of me would be wrong. Right?
2) So I think when tau varies exponentially, as in Erlang flows, then already these ACF values can be trusted...
3) Sorry, maybe I'm being silly, but I badly need a parameter that calculates process "memory". I'm really hoping for ACF...
1) Right.
2) This is nonsense.
3) Yes, nonsense... You still don't understand what ACF is. Anyway, ACF is not"a parameter that calculates the 'memory' of a process".
.
Judging by the comments on the forum about ACF, I come to the conclusion that for many the degree of understanding of ACF, to put it mildly, is close to zero.
Apparently, it would not be a bad idea to take ACF apart into its bones, to break it down. So that it would be clear what is connected to what and where it leads.
And more broadly, to cover the correlation analysis. And then spectral analysis as well.
This could make a series of articles.
However, I doubt it...Judging by the comments on the forum about ACF, I come to the conclusion that for many the degree of understanding of ACF, to put it mildly, is close to zero.
Apparently, it would not be a bad idea to take ACF apart into its bones, to break it down. So that it would be clear what is connected to what and where it leads.
And more broadly, to cover the correlation analysis. And then spectral analysis as well.
That could make for a series of articles.
However, I doubt it...Why not? They say they pay good money for articles - don't keep running a free branch all the time. IMHO.
I am personally interested in two questions about ACF:
1. is it possible to change the tau shift exponentially at each calculation step, in a sliding window?
2. will the correlation decrease with increasing tau?
If the articles give reasoned answers, without the emotion inherent in you and me, I would be very grateful.
P.S. All my posts are by no means flattery etc, Oleg. I really need help now, in the face of such a serious adversary as the market. Well, I can't embrace the immensity alone - and there are no voluntary helpers here and never has been, apparently...
Judging by the comments on the forum about ACF, I come to the conclusion that for many people the degree of understanding of ACF is, to put it mildly, close to zero.
Apparently, it would not be a bad idea to take ACF apart into its bones, to break it down. So that it would be clear what is connected to what and where it leads.
And more broadly, to cover the correlation analysis. And then spectral analysis as well.
That could make a series of articles.
However, I doubt it...It doesn't make sense, imho.
They should just read a textbook. It's all laid out before you. And well laid out. It's all available. It is only a question of desire.
Imho, it doesn't make sense.
They should just read a textbook. It's all laid out before you. And well laid out. It's all available. It's only a question of desire.
I'm inclined to think there's no point, either.
That's true... But here's an example of total incomprehension, with questions for which you won't find direct answers (either yes/no or detailed and with arguments) in a textbook:
I am personally interested in two questions about ACF:
1. Is it possible to change the tau shift exponentially at each computational step, in a sliding window?
2. will the correlation decrease with increasing tau?
If the articles give reasoned answers, without the emotion inherent in you and me, I would be very grateful.
P.S. All my posts are by no means flattery etc, Oleg. I really need help now, in the face of such a serious adversary as the market. Well, I can't embrace the immensity alone - and there are no volunteers here, nor have there ever been, judging by all appearances...
what can I say...
1. you can, but it makes no sense at all
2. it may either decrease or increase -- it depends on the process in question (BP)