Adaptive digital filters - page 22
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
has anyone made an adaptive filter using the Herzel algorithm, it was even posted in mql on the forum (attached) on its basis it seems to me it is possible to make an adaptive harmonic (adaptive filter) without using a digital filter generator
What's good on the forum - as you open it you always come to grips with it. There are no signals in the marketplace, you see, well there are just no signals, so there is simply nothing to adapt to.I am not talking about signals but harmonics (and how to use them in TS depends on a particular TS) or a kind of oscillator of harmonics and how to tune a particular harmonic so that it filters out unwanted fluctuations within certain limits
There are no harmonics, that's the problem. I've posted BP spectra with a shift of a few bars on the forum several times.
Что хорошо на форуме - как откроешь всегда приходишь в умиление. На рынкете нет сигналов, понимаете, ну просто нет сигналов, поэтому адаптироваться просто не к чему.More accurately, the quote is the signal itself. Correct filtering will give practically (indistinguishably the same) original signal. The term "correct" means that after subtracting the filter from the initial signal there will be noise, and proven noise, and not something else.
And from the point of view of "the physics of the phenomenon" filtering of a quote (exactly a quote) does not make much sense, unlike other applications in TS. When, for example, a multi-ton space vehicle is docked in space, of course, this machine must be controlled. And to do that, a bunch of parameters need to be controlled. And I think that practically all the parameters are measured with errors, and not only because of noise, but maybe as a result of some environmental influences. Filtering is used to estimate true value (e.g. speed). After all, it is not easy to control without knowing, for example, velocity.
The magnitude of errors in "measurements" of quotes of brokerage companies and their "distribution" is almost negligible (of course, there are very rare cases, but the recovery of "failures" is a separate procedure, no filters will help). And the DC will not make a deal at a price that is not really available anywhere :o)
Да нет гармоник, в этом проблема. Я на форуме несколько раз выкладывал спектры ВР со сдвигом несколько баров.There are always harmonics. Simply using the Fourier transform also makes no sense. The thing carries no information for the quotation process.
There are always harmonics. Simply using the Fourier transform also makes no sense. This stuff doesn't carry any information for the quoting process.
I clarify: there are always harmonics, but they usually do not live long and in general their lifetime is indefinite. You find a harmonic, and it dies or doesn't die. Which one are we talking about, the one that died or the one that didn't die?
More accurately, the quote is the signal itself. Correct filtering will give practically (indistinguishably the same) original signal. The term "correct" means that after subtracting the filter from the initial signal there will be noise, and proven noise, and not something else.
And from the point of view of "the physics of the phenomenon" filtering of a quote (exactly a quote) does not make much sense, unlike other applications in TS. When, for example, a multi-ton vehicle is docked in space, of course, we need to control this machine. And to do that, a bunch of parameters need to be controlled. And I think that practically all the parameters are measured with errors, and not only because of noise, but maybe as a result of some environmental influences. Filtering is used to estimate true value (e.g. speed). After all, it is not easy to control without knowing, for example, velocity.
The magnitude of errors in "measurements" of quotes of brokerage companies and their "distribution" is almost negligible (of course, there are very rare cases, but the recovery of "failures" is a separate procedure, no filters will help). And the DC will not make a deal at a price that really does not exist anywhere :o)
I write about DSP. Radio, television, location, etc. There is always a signal source and we have a certain, a priori knowledge of that signal. If the price is a signal, what is it about? A sequence of prices, BP, should give us a signal about a position in the market. No separation of noise gives a market position, there are no algorithms, no mathematics for that. The classical approach is to identify BP (get some parameters) and having identified BP by algorithm get a position. I have seen an article where it is proved that it is impossible to identify the trend as such.