Discussing the article: "Discrete Hartley transform"

New comment
 

Check out the new article: Discrete Hartley transform.

In this article, we will consider one of the methods of spectral analysis and signal processing - the discrete Hartley transform. It allows filtering signals, analyzing their spectrum and much more. The capabilities of DHT are no less than those of the discrete Fourier transform. However, unlike DFT, DHT uses only real numbers, which makes it more convenient for implementation in practice, and the results of its application are more visual.

In 1942, Ralph Hartley proposed an analogue of the Fourier transform in his article "A More Symmetrical Fourier Analysis Applied to Transmission Problems".

Just like Fourier transform (FT), Hartley transform (HT) turns the original signal into a sum of trigonometric functions. But there is one significant difference between them. FT converts real values to complex numbers, while HT provides only real results. Because of this difference, the Hartley transform did not become popular - scientists and technicians did not see any advantages in it and continued to use the usual Fourier transform. In 1983, Ronald Bracewell presented a discrete version of the Hartley transform.

Author: Aleksej Poljakov

 
Thanks for the interesting article! Some things look very promising. :)
 
Really enjoyed the article. Very similar to what can be applied in practice. I downloaded the indicators, used them on different timeframes. So far, I got the impression that they give effective signals. I will test more.
 

Thanks for the article!

Just yesterday I was thinking about methods of identifying market stages through statistical indicators in a subsample, and I saw similar ideas in the article.

Have you done any deeper research in this direction? I wonder how fast (with what lag) and with what error it was possible to classify trends\flats?

 
Aleksey Vyazmikin #:

Thanks for the article!

Just yesterday I was thinking about methods of identifying market stages through statistical indicators in a subsample, and I saw similar ideas in the article.

Have you done any more in-depth research in this direction? I wonder how fast (with what lag) and with what error it was possible to classify trends\flats?

Identifying market stages is probably the easiest task. We take the price spectrum (without the main signal) and pass it through cluster analysis (Kohonen maps) - that's what we get market stages. But everything is very complicated with trends - the sign of change/beginning of a new trend is relative weakness of the low-frequency component and relative strengthening of high-frequency harmonics. But, unfortunately, it is possible to seriously miss the trend direction.

New comment