Quantum mechanical methods - page 7

New comment
 
Lo083:
Does the filter there have anything to do with uncertainty ? I.e. is it written somewhere in the book that it is used there in the construction of the filter ?
Any filter is by definition limited by the fundamental problem mentioned. Try to synthesize a non-delayed filter :-0 Though - try to justify the fundamental prohibition on its construction? That's the trick, there are no fundamental bans.
 
Dr.Fx:
Funny how people discover that mechanics equations are covariant with respect to the time inversion operation. why wouldn't they be covariant? energetically everything satisfies the principle of least action on the whole trajectory, which is the only one.

Comrades, can we keep it simple for "dummies"? For example. In what area of thechanalysis can the laws of quantum mechanics be applied? 2015.03.20 07:07:31

Topikstarter: links to the website can be inserted directly. For example. https://www. mql5.com/ru/forum/114318
Квантовая торговая система - MQL4 форум
  • www.mql5.com
Квантовая торговая система - MQL4 форум
 
Dr.Fx:
Any filter is by definition limited by the fundamental problem mentioned. Try to synthesize a non-redundant filter :-0 Though - try to justify a fundamental prohibition on its construction? That's the trick, there are no fundamental prohibitions.
Filter cannot be non-lagging, it shows some frequency when it is present, when it is absent, it does not show it. Or ? Uncertainty seems to me to apply to filters in a broader sense.

 
Yuri_Evseenkov:

Comrades, can we keep it simple for "dummies"? For example. In what area of thechanalysis can the laws of quantum mechanics be applied? 2015.03.20 07:07:31

Topikstarter: links to the website can be inserted directly. For example. https://www. mql5.com/ru/forum/114318
It does not apply to "technical analysis". There is another analysis being discussed here. There is a website (link will be sent to personal forum email) if participants meet, there can be a more detailed discussion.
 
Lo083:
The filter cannot be non-lagging, it shows some frequency when it is there, when it is not, it does not. Or ? Uncertainty seems to me to apply to filters in a broader sense.

You should understand a little bit what a digital filter is... then you wouldn't be talking about "the frequency it shows" :-)
 
Dr.Fx:
You should understand a bit what a digital filter is... Then you wouldn't be talking about "the frequency it shows" :-)
I don't understand how you define frequencies which must be selected by the filter. I understand, that e.g. the Fourier filter doesn't have it's own parameters. Manually set filter parameters and guess on paper? :-) Filters are a different story, but it's interesting that you found them in vagueness, if it's not a mistake. :-)

It seems to me that these parameters should not be set "manually", but should be determined by "what is in the time series".

//---- input parameters of the indicator
input FType_ FType = LPF; //Filter type
//0 - lowpass (FATL/SATL/KGLP), 1 - highpass (KGHP), 2 - bandpass (RBCI/KGBP), 3 - regenerative (KGBS)
input int P1 = 28; //Drop period 1, bar
input int D1 = 19; //transient cut-off period 1, bar
input int A1 = 40; //Damping within delay band 1, dB
input int P2 = 0; //transient cut-off period 2, bar
input int D2 = 0; //transient cut-off period 2, bar
input int A2 = 0; //Delay in the delay band 2, dB
input int Delay = 0; //Delay, bar
input double Ripple = 0.08; //Bandwidth attenuation, dB
input int FILTERShift = 0; //mux shift by horizontal bar.

I haven't seen any filters, which don't specify frequency, most likely don't exist :-)
 
Lo083:
I do not understand how you define the frequencies that should be selected by the filter, as I understand that, for example, there is no Fourier filter for selecting the required filter parameters. Do you set filter parameters manually and estimate them on paper? :-) Filters are a different story, but it's interesting that you found them in vagueness, if it's not a mistake. :-)
I have no idea which "frequencies" you are talking about. But if you need to decompose a signal into some frequencies, keep in mind that the FFT doesn't do that: the Fourier transform breaks the signal not into HIS frequencies (the "true" frequency spectrum inherent in it), but into TOTAL (pre-defined by a rigid transformation grid stemming from the number of samples to be transformed. In short, the result is a property of the transform, not of the original object. An example? Take a 100 Hz sine wave. Subject it to Fourier transforms. See in the spectrum (on a finite number of samples) anything but a stick at 100 Hz. Why this 100 Hz sine wave spectrum is thought to contain anything other than 100 Hz is the big problem of Fourier analysis.
 
Dr.Fx:
I have no idea what "frequencies" you are talking about. But if you need to decompose a signal into some frequencies, keep in mind that the FFT does not do that: the Fourier transform does not split the signal into HIS (inherent, in some way "true" frequency spectrum) frequencies, but into HIS (predetermined by a rigid transformation grid stemming from the number of samples to be transformed. In short, the result is a property of the transform, not of the original object. An example? Take a 100 Hz sine wave. Subject it to Fourier transforms. See in the spectrum (on a finite number of samples) anything but a stick at 100 Hz. Why it is thought that the spectrum of that 100 Hz sine wave contains anything but 100 Hz is the big problem of Fourier analysis.
No, Fourier is used to see the frequency with the maximum amplitude present in the series. I've heard there are not only frequency filters, but also others, but I haven't studied which ones.
 
Lo083:
It seems to me that these parameters should not be set "manually", but should be determined by "what is in the time series".

//---- input parameters of the indicator
input FType_ FType = LPF; //Filter type
//0 - low pass filter (FATL/SATL/KGLP), 1 - high pass filter (KGHP), 2 - bandpass filter (RBCI/KGBP), 3 - regenerative filter (KGBS)
You are in the middle of nowhere. You have nothing to choose from. All filters described in literature are linear or adaptive. What is adaptive - it is also linear, but a little non-stationary - with slowly changing (specify, of course, relative to what slowly) parameters. There is a theorem forbidding a linear filter from lagging. That's it. Deadlock. The problem under discussion cannot even be put in known terms. For there is no prohibition for creating a nonlinear non lagging (not looking into the future that would violate the principle of causality, but simply non lagging that does not contradict anything). And there never has been. But where to start the synthesis of such a filter cannot be clear to you. For all known theory of filtering should be thrown out. There is no definition even of concept "filter" in it. There is no definition of "lag". There is no AFC/frequency, no FIR, no BIH, nothing. :-)
 
Lo083:
No, Fourier is used to see the frequency with the maximum amplitude present in the series,
Once again: Fourier does not in principle give frequencies "present in the series". It gives an approximation to a DESIGNED frequency grid. This is a typical manifestation of a fundamental principle of measurement theory: what we end up observing is not a property of the object, but a convolution of the object and probe (instrument, or, in this case, algorithm) properties.
12345678910111213
New comment