Spectrum derivative (or spectrum acceleration) - page 2

 
Reshetov:
Well, if BP is not periodic or the period is not correct, it is bound to change.

It is like a signal in the form of a sine wave with irregularly torn out pieces of data, you can not recover the signal and see the periodicity without the missing pieces, for this you need to use a multi-currency (cluster) analysis - to look for periodicity and inertia within the cluster. but for now, not about that, to understand how to properly and logically decompose each series (pair) separately, so that then use in the form that wrote in the preceding sentence.
 

And something similar is obtained for each of the frequencies. Amplitude at the top, phase at the bottom. Another thing to consider is that this is a period delayed indicator (for phase), in this case 10 bars.

 

This is the amplitude. 1000 counts, max period 200.

 
trollolo:

Actually the title of the topic speaks for itself.

Start here https://www.mql5.com/ru/forum/137042, interested in who researched and how deep this question is, also interested in phase observations in this case.

"Spectrum derivative (or spectrum acceleration)" is a good topic for schoolchildren. Have you, ever, calculated a spectrum? Do you know how to do it?

The simplest one is a power series. In this series (by the number of the row term): 0- displacement, 1-velocity, 2-acceleration, 3-acceleration of acceleration, 4-acceleration of acceleration, and so on. Question: What are you talking about?

Speed and acceleration was the only good topic in November 2009.

 
Rorschach:

This is the amplitude. 1000 counts, max period 200.


So? What does that tell you? And why the max period is only 200 and not more.

very interesting how you get them so beautiful with Ilya)) or here https://www.mql5.com/ru/forum/118473

 
icas:

"The derivative of a spectrum (or spectrum acceleration)" is a good topic for schoolchildren. Have you ever calculated a spectrum? Do you know how to do it?

The simplest one is a power series. In this series (by the number of the row term): 0- displacement, 1-velocity, 2-acceleration, 3-acceleration of acceleration, 4-acceleration of acceleration, and so on. Question: What are you talking about?

Speed and acceleration was the only good topic in November 2009.

ask for the link.
 
Rorschach:

And something similar is obtained for each of the frequencies. Amplitude at the top, phase at the bottom. Another thing to consider is that this is a period delayed indicator (for phase), in this case 10 bars.


You probably have a slightly different Hertzl algorithm, I found a decomposition using it on the forum, it seems to have more than one frequency decomposition

 
trollolo:


and ? what does it say? and why the max period is only 200 and not more.

very interesting how you get them so beautiful with Ilya)) or here https://www.mql5.com/ru/forum/118473


and ? what does it say? I was expecting to see a more pronounced periodicity, as here for example. It's also interesting that there are no intervals between frequencies (periods) like there. I don't know what else to add yet, I'll keep digging.

And why the max period is only 200 and not more. Matcad refused to work with a bigger file(((?

 
trollolo:


Your algorithm seems to have a slightly different Hertzl algorithm, I found a decomposition using it on the forum, it seems to have more than one frequency decomposition


I had a look at this indicator but did not get what it was counting. I'll show you mine now. My calculations are taken from here. The examples also explain how to select the number of bars and spectral count.
Files:
 
trollolo:

Actually the title of the topic speaks for itself.

Started here https://www.mql5.com/ru/forum/137042, interested in who researched and how deep this question is, also interested in phase observations in this case.


I think I'm beginning to get the hang of it.

The McDi formula is a bandpass filter, the output is the sum of the frequencies (what is displayed as an indicator). So you want to calculate the future value of this sum using velocity and acceleration (1 and 2 derivatives).

The macdi from macdi from..... - is an N-order filter, it improves filtering but increases time lag from price (it's more efficient to make macdi based on JJMA).

I found something interesting about phase, but I don't fully understand it. Negative GVZ, for example here and here. As far as I understand it has the effect of predicting certain frequencies.