Discussion of article "Introduction to the Empirical Mode Decomposition Method"

New comment
 

New article Introduction to the Empirical Mode Decomposition Method is published:

This article serves to familiarize the reader with the empirical mode decomposition (EMD) method. It is the fundamental part of the Hilbert–Huang transform and is intended for analyzing data from nonstationary and nonlinear processes. This article also features a possible software implementation of this method along with a brief consideration of its peculiarities and gives some simple examples of its use.

Empirical Mode Decomposition MQL5

Author: Victor

 
MetaQuotes:

Published article Familiarity with the method of empirical mode decomposition:

Author: Victor

Good material. Too bad the Hilbert-Huang transformations are not considered.
 
HideYourRichess:
Good material. It is a pity that Hilbert-Huang transformations are not considered.

Somehow it so happened that I was interested only in the first part of HHT, i.e. EMD. I wanted to feel the results of decomposition. As for the next step, the Hilbert transform, I just didn't get to it. Probably because there was no need to construct the instantaneous spectrum. So, I did not even try to supplement EMD with Hilbert transform.

 

In general, yes, you can do without spectra, especially if one EMD is enough for practice.

 

1. It is very symptomatic that one of the first decompositions into components by Box and Jenkins, known as ARMA, is named and not one of the first decompositions into components. Naturally, ARCH is not mentioned.

2. It would be possible to omit mentioning the said very respected citizens in the world, if it were not for one thing: they described the problem and from the problem they formulated the purpose of the decomposition. From these premises, both the limitations of the method and the price of the result become clear.

3. What is the purpose of all this? What problems does the decomposition solve? What are the limitations? What remains unsolved?

4. Most importantly: does this decomposition have the property of predictability?

5. Judging by the overwhelming number of the author's articles, he came to the market from the DSP, which is centred on analysis with a very specific goal: to identify a signal from the noise. There is no signal on the market in the sense of DSP, but there are trends, so we look for them. In analysis we try to find a trend in the hope of its continuation. That's why analysis is interesting in trading only as a forecast substantiation - analysis itself is not interesting. Any change of position to the market "out of the market - in the market" is a decision made on the basis of a forecast of the near future.

 
faa1947:

5. Judging by the overwhelming number of the author's articles - he came to the market from the DSP, . . .

Actually, I came to the market from the Multiplication Table.

 
You should write an Expert Advisor or Expert Advisor for MQL5, of course with constant updates, since you yourself indicated in the article that this method has not yet been fully explored. Although for the beginning you can try the script, how it goes, and then you will see.
 
victorg:

I actually came to the market from the Multiplication Table.

We all come from childhood.

You constantly walk around the huge and mighty building"econometrics (statistics)", picking out separate pieces or bricks from it, persistently not noticing the whole knowledge and not trying to attach your research as one of its bricks. Because of this your articles hang in the air, although they are undoubtedly interesting both in terms of level and subject matter.

 

Specifically on the article.

You have highlighted several components. Which of them or their combination has value for trading and why?

 
faa1947:

We all come from childhood.

You constantly walk around the huge and mighty building "econometrics (statistics)", picking out separate pieces or bricks from it, persistently not noticing the whole knowledge and not trying to attach your research as one of its bricks. Because of this your articles hang in the air, although they are undoubtedly interesting both in terms of level and subject matter.

This is rather a fundamental problem of our time. The amount of information in every field of knowledge has become so large and the models so complex that it has become difficult for one person to cover any science in its entirety. Therefore, we have to deal only with specific sections of a particular science in order to have time and be able to understand something. Imho! :)
 
-Alexey-:
This is rather a fundamental problem of our time. The amount of information in every field of knowledge has become so large and the models so complex that it has become difficult for one person to cover any science in its entirety. Therefore, we have to deal only with specific sections of a particular science in order to have time and be able to understand something. Imho! :)

There is no question of encyclopaedic knowledge of econometrics. It is enough to select something acceptable and use it - buy a car and drive. Moreover, in this discussion we are talking about a very narrow issue: decomposition of a quotient into its components. The classical decomposition in the last 50 years is the Box and Jenkins decomposition: trend + cycle + noise (residual of the sum of the first two). This decomposition makes economic sense. Wavelets are mentioned - they give a different decomposition, but it also makes economic sense and wavelets fit well with Box's idea.

What do we have in the article? This is exactly the question I am interested in.

12345
New comment