create an expert for mt4 using a programme made in exel - page 23

 
Gentlemen, will the report become available to mere mortals?
 
Sorento:

If alsu could tell me about his approximations by exponentially damped cosines, I'd be very interested in this

maybe this:


http://www.google.ru/search?hl=ru&source=hp&q=vjuvers&aq=f&aqi=&aql=&oq=

 
vasya_vasya:
Gentlemen, will the report become available to mere mortals?
The paper is being prepared for publication. There are a lot of formulas that need to be put into the right form. It takes time.
 
Mathemat:
The article is being prepared for publication. There are a lot of formulas which need to be put into the right form. It takes time.

Miracles. And what is she going to popularise?

MQL5 4?

Or their future users?

;)

 
Sorento:

If alsu had told me about his approximations by exponentially attenuated cosines, I would have been more interested.

And they are not mine, they are Laplace's.)

If you want to discuss it, I'll give you the premise. In application to a series with discrete time the Laplace transform is not used in its pure form, it is reduced to so-called Z-transformation, and they are translated to each other by simple replacement z = exp(s*T), where T is a sampling period. So, the damped (and not only divergent) sine-cosines are obtained when we perform inverse transformation from z- (or s-) domain to time domain: in doing so we have to perform integration over a contour on the complex plane covering the convergence domain and all image poles (there is a mistake in wikipedia - it says "covering subtractions"). Just on this closed contour, because z will take values with different real and imaginary parts, our sine-cosines emerge: the real part of the exponent, recall, corresponds to damping parameter (or divergence, if it is positive), imaginary part to circular frequency. We obtain approximately the same principle as in the Fourier transform - only the exponent exponents do not have a real part there. Thus, the Z-transform is a generalization of the discrete Fourier transform, and the latter is obtained from Z by choosing the unit circle z = exp(jw) as the integration contour.

I hope you are familiar with complex analysis, otherwise it would be difficult to explain...

 
alsu:

And they're not mine, they're Laplace's.)

If you want to discuss it, I'll give you a message. In application to a series with discrete time the Laplace transform is not used in its pure form, it is reduced to so-called Z-transformation, and they are translated to each other by simple replacement z = exp(s*T), where T is sampling period. So, the damped (and not only divergent) sine-cosines are obtained when we perform inverse transformation from z- (or s-) domain to time domain: in doing so we have to perform integration over a contour on the complex plane covering the convergence domain and all image poles (there is a mistake in wikipedia - it says "covering subtractions"). Just on this closed contour, because z will take values with different real and imaginary parts, our sine-cosines emerge: the real part of the exponent, recall, corresponds to damping parameter (or divergence, if it is positive), imaginary part to circular frequency. We obtain approximately the same principle as in the Fourier transform - only the exponent exponents do not have a real part there. So, Z-transform is a generalization of discrete Fourier-transform and the latter is obtained from Z by choosing unit circle z = exp(jw) as integration contour.

I hope you are familiar with complex analysis, otherwise it would be a bit difficult to explain...

Thanks))

I was actually talking about the practical part, as it were, about the results and obstacles...

 
Mathemat:
The article is being prepared for publication. There are a lot of formulas that need to be put in the right form. This takes time.
Yes, there will be a lot of formulas.
Files:
30.rar  2160 kb
 
Sorento: Miracles. And what will it popularise?
Apparently, it will popularise the gamma function and the corresponding probability distribution:)
 
Sorento:

Thank you.))

I was actually talking about the practical part, as it were, about the results and obstacles...

Well, the obstacles here are the same as in conventional discrete Fourier - windows, spectrum overlaps, resolution... the results are better because the functions are asymptotically converging to zero.
 
Richie:
Yes, there will be many formulas.

:)))

what are the lyrics, and what musical is the song from ?