Not who, but what - understanding the simple fact that we are dealing with a sample value that only makes sense (converges to something as the sample size increases) when it is stationary.
Not who, but what - an understanding of the simple fact that we are dealing with a sample value that only makes sense (converges to something as the sample size increases) when it is stationary.
The ACF makes sense for any signal, even a non-stationary one.
Andrei: ACF makes sense for any signal, even a non-stationary one.
No. For non-stationary, it makes sense to talk about a QF that will depend on two variables, not one, like the ACF. It is possible to somehow (in a large number of different ways) make this QF into something that depends on one variable. But don't call it an ACF - don't confuse yourself and others.
No. For non-stationary, it makes sense to talk about a QF that will depend on two variables, not one, like the ACF. It is possible to somehow (in a large number of different ways) make this QF into something that depends on one variable. But you shouldn't call it an ACF - no need to confuse yourself and others.
Not really. ACF in this case is simply the classical convolution of any signal on some limited segment with its copy.
There is nothing unusual about it, nor is there any reason to panic.
How many variables the ACF depends on does not play a role here.
(converges to something with increasing sample size) only at stationarity.
everything converges to something, whatever you count, there is number theory, even it has regularities, which appear with a large sample of values, although it (number theory) does not study any physical or other process
A necessity of the autocorrelation function for more than one parameter was mentioned in the thread, it is a research from the field of fields, I doubt that a discrete function in the time scale (price series) is worth considering by a field
And in general, correlation analysis of a non-periodic function, what should it show? Correlation analysis in a periodic function will show the frequency spectrum distribution, and what should the price chart correlation analysis show?
I found a good read on the subject, very similar to the textbook I studied 20 years ago http://scask.ru/book_brts.php?id=16
Изучая АКФ пачки прямоугольных видеоимпульсов, читатель, безусловно, обратил внимание на то, что соответствующий график имел специфический лепестковый вид. С практической точки зрения, имея в виду использование АКФ для решения задачи обнаружения такого сигнала или измерения его параметров, совершенно несущественно, что отдельные лепестки имеют...
everything converges to something, whatever you count, there is number theory, even it has regularities that appear in a large sample of values, although it (number theory) does not study any physical or other process
A necessity of the autocorrelation function for more than one parameter was mentioned in the thread, it is a research from the field of fields, I doubt that a discrete function on the time scale (price series) may be considered by a field
And in general, correlation analysis of a non-periodic function, what should it show? Correlation analysis in a periodic function will show the frequency spectrum distribution, and what should the price chart correlation analysis show?
We need a measure of "memory" - a specific numerical value of the dependence of price increments on each other in the time sliding window.
This makes it possible to say whether the sum of the increments in that window forms a number belonging to the Gaussian distribution or not.
In fact, the ACF is the Grail, folks! It shows whether we are in a trend or flat area...
You just have to learn how to calculate it correctly - that's what I'm doing now...
Allow the use of cookies to log in to the MQL5.com website.
Please enable the necessary setting in your browser, otherwise you will not be able to log in.
You are signed in. Two-factor authentication has been disabled. Set it up again and scan a new QR code, or continue with two-factor authentication disabled.
not really, allows to separate periodic components...
autocorrelation, as well as correlation, are not applicable to non-stationary times
autocorrelation as well as correlation are not applicable to non-stationary times
Who's stopping you from using it?
Overestimation, Spearman is no better.
Over-rating comes out, Spearman is no better either
and who prevents it from being applied?
Not who, but what - understanding the simple fact that we are dealing with a sample value that only makes sense (converges to something as the sample size increases) when it is stationary.
Not who, but what - an understanding of the simple fact that we are dealing with a sample value that only makes sense (converges to something as the sample size increases) when it is stationary.
ACF makes sense for any signal, even a non-stationary one.
No. For non-stationary, it makes sense to talk about a QF that will depend on two variables, not one, like the ACF. It is possible to somehow (in a large number of different ways) make this QF into something that depends on one variable. But don't call it an ACF - don't confuse yourself and others.
No. For non-stationary, it makes sense to talk about a QF that will depend on two variables, not one, like the ACF. It is possible to somehow (in a large number of different ways) make this QF into something that depends on one variable. But you shouldn't call it an ACF - no need to confuse yourself and others.
Not really. ACF in this case is simply the classical convolution of any signal on some limited segment with its copy.
There is nothing unusual about it, nor is there any reason to panic.
How many variables the ACF depends on does not play a role here.
(converges to something with increasing sample size) only at stationarity.
everything converges to something, whatever you count, there is number theory, even it has regularities, which appear with a large sample of values, although it (number theory) does not study any physical or other process
A necessity of the autocorrelation function for more than one parameter was mentioned in the thread, it is a research from the field of fields, I doubt that a discrete function in the time scale (price series) is worth considering by a field
And in general, correlation analysis of a non-periodic function, what should it show? Correlation analysis in a periodic function will show the frequency spectrum distribution, and what should the price chart correlation analysis show?
I found a good read on the subject, very similar to the textbook I studied 20 years ago http://scask.ru/book_brts.php?id=16
everything converges to something, whatever you count, there is number theory, even it has regularities that appear in a large sample of values, although it (number theory) does not study any physical or other process
A necessity of the autocorrelation function for more than one parameter was mentioned in the thread, it is a research from the field of fields, I doubt that a discrete function on the time scale (price series) may be considered by a field
And in general, correlation analysis of a non-periodic function, what should it show? Correlation analysis in a periodic function will show the frequency spectrum distribution, and what should the price chart correlation analysis show?
We need a measure of "memory" - a specific numerical value of the dependence of price increments on each other in the time sliding window.
This makes it possible to say whether the sum of the increments in that window forms a number belonging to the Gaussian distribution or not.
In fact, the ACF is the Grail, folks! It shows whether we are in a trend or flat area...
You just have to learn how to calculate it correctly - that's what I'm doing now...