Adaptive digital filters - page 6

New comment
 
Greetings esteemed assembly.
grasn:

  • Collect statistics on "waves" (wavelengths/channels bounded by extrema, sweep...) based on VLF
  • Find statistical patterns between previous and future waves. By wave, we mean the channel that limits the wave, i.e. not a straight line, but a time line

Grasn, how labor-intensive for you to carry out such a Data Mining on the statistics of others? Question, understandably with a hint :).

Concerning the identification of extrema in filtered data I think I have already written that the market respects levels, therefore it is more preferable to identify Highs and lows by High and Low. Especially taking into account the information lost when compressing it into a bar chart. False extrema create problems, but I am afraid the LFO, even an adaptive one, will not always help. Below is an example of extrema that look false:


The red bars here are my ideas about the correct markup(horizontal lines are added as an argument in favor of this opinion).
 
Candid, I would argue with you. Your argument is not convincing. These "false" levels can also come from higher TF as a result of clustering of several Fibs from different swings. I have no proof, it's only a hypothesis.
 
grasn:

PS1: I served just in the Air Defense Forces, here I am reading out my position from paragraph 27 of my military ID (to make it more solid :o): "commander of the department of anti-aircraft short-range missile radio control facilities". And I know more than well (as they usually write - not by hearsay :o) that our praised systems (and not only ours) cannot even shoot down, they cannot even SEE the targets.


And why are they so afraid, I think it would be interesting for you http://www.kroufr.ru/forum/index.php/topic,6037.0.html, and this SAM (development) is more than 50 years old, so they see and not bad.
 
Mathemat:

Integer, you mean this JMA - 'JMA'?


About her.
 
NorthernWind:

ZS, on the Alpari forum, BQQ has laid out in some detail why, in his opinion, as a DSP specialist, DSP methods are difficult to apply in the market. Quite lucid, in my opinion.


I BQQ argued a little here http://forum.alpari-idc.ru/showthread.php?t=38804&page=16, if it's not too much trouble, where he lays out all the details. I just think that first of all a DSP specialist should know and understand (with all his soul) this theorem of Kotelnikov, it is like an axiom in geometry.

And to all if you can use the term sample rate please, to me the Neukvist frequency is a dirty word. It's from the area who invented radio Popov or Marconi, etc.

to Integer

I've been working on JMA for two days and it's hopeless!

If you don't mind, you can try to make an adaptive indicator, the algorithm I wrote earlier in this thread. If you want to try it, you've got to know how to work with JMA.

 
Integer:
About her.
Are you sure this is the original JMA? It's just thatParabellum posted a picture of the JJMA in the discussion, which seems to be better...
 
Prival:
Something about the JMA, like the best, adaptive, etc., hit me. (all eaten up, how). And we have a good job :-). And the left-handed like Russia is no more, but I do not believe it.
I look, look at him - some strange formulas, and the avatar is not something like :-) I like it better :-).
(Compare http://www.jurikres.com/catalog/ms_ama.htm#top). Our aeroplane is better :-).

That`s why I suggest to try to make a better indicator, more adaptive. Maybe something good will come out.

The idea is the following.
1. We take this indicator as a base ('Kaufman optimized AMA: Perry Kaufman AMA optimized'), many people have already worked on it. The theory of this indicator is described in the file (file attached). We take one part of this indicator (idea). Calculation of the ER efficiency ratio (varies from 0 to 1). It will determine the averaging (sampling) period from 2 to N (N is set as an input parameter in the algorithm). The rest is a bit trickier.
2. we do not use EMA (exponential moving average) but a polynomial. the maximum degree of the polynomial is n (also set as an external parameter). we can stop and vary n and run it in the tester, I think we can already get good results. But IHMO the flea is not yet fully trained, so let's move on.
3. If it's adaptive, then let it be adaptive to the fullest extent. In addition, the next one - the degree of polynomial is also calculated (chosen the best one by some criterion). Since we have no a priori information on noise. I suggest using the criterion - the coefficient of determination. The logic of selecting the optimal polynomial according to this criterion is described in the file (see pp. 12, 13 and 14). There is even a program written in MathCade, how to do it.

If anyone is interested, I am ready to program and recheck point 3 in MathCade. I will also help with creation of such indicator in MQL due to my modest capabilities.

Prival, this archive does not contain page 12-13
 
Mathemat:
Integer:
About her.
Are you sure this is the original JMA? It's just thatParabellum posted a picture of the JJMA in the discussion, which seems to be better...

I'm not sure.
 
Mathemat:
Candid, I would argue with you. Your argument is not convincing. These "false" levels can also come from higher TF as a result of clustering of several Fibs from different swings. I have no proof, it is only a hypothesis.
No, Mathemat, I'm not going to argue with you about that :). Because in principle I agree. But I think it is highly desirable to reduce the problem to an independent search for patterns for each rank (I prefer the notion of "higher timeframe" to the notion of "higher rank"). But in general the idea that we are dealing with some kind of interference pattern looks interesting.
 
Integer:
Prival:
Something hit me about JMA, like the best, adaptive, etc. (he ate them all up). But we are not good at it :-). And the left-handed like Russia has become extinct, but I do not believe it.
I look, look at him - some strange formulas, and the avatar is not something like :-) I like it better :-).
(Compare http://www.jurikres.com/catalog/ms_ama.htm#top). Our aeroplane is better :-).

That`s why I suggest to try to make a better indicator, more adaptive. Maybe something good will come out.

The idea is the following.
1. We take this indicator as a base ('Kaufman optimized AMA: Perry Kaufman AMA optimized'), many people have already worked on it. The theory of this indicator is described in the file (file attached). We take one part of this indicator (idea). Calculation of the ER efficiency ratio (varies from 0 to 1). It will determine the averaging (sampling) period from 2 to N (N is set as an input parameter in the algorithm). The rest is a bit trickier.
2. we do not use EMA (exponential moving average) but a polynomial. the maximum degree of polynomial n (also set as an external parameter). we can stop and vary n and run it in the tester, I think we can already get good results. But IHMO the flea is not yet fully trained, so let's move on.
3. If it's adaptive, then let it be adaptive to the fullest extent. In addition, the next one - the degree of polynomial is also calculated (chosen the best one by some criterion). Since we have no a priori information on noise. I suggest using the criterion - the coefficient of determination. The logic of selecting the optimal polynomial according to this criterion is described in the file (see pp. 12, 13 and 14). There is even a program written in MathCade, how to do it.

If anyone is interested, I am ready to program and recheck point 3 in MathCade. I will also help you to create such indicator in MQL due to my modest capabilities.

Prival, this archive does not contain pages 12-13
Ok, sorry, I will add them again (topic 14, Approximation of Signals, pp. 12-14). But the 3rd item, I think, is not necessary, it is possible for the beginning simply to choose polynomial of 1st or 2nd degree. Since for the third point you need to answer 1, but the most important question, what is the regular component here (split into signal and noise)
Files:
dsp.zip  1921 kb
12345678910111213...41
New comment
Reason: