Machine learning in trading: theory, models, practice and algo-trading - page 3687
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
Thanks, I'll look at it later.
There's not much to see, no details, just facts.
Forext is a real system, it has been used in practice, I have seen, though long ago, something calculated for the US government.
But I have never heard of any Fourier system that anyone has used successfully.
Forexst is a real system, used in practice, I saw, though long ago, something counted for the US government.
But I have never heard of any Fourier system that anyone has used successfully.
It is necessary to go through the variants, maybe I will look at this one later, if the topic is interesting. I'm tired of marking marks, I want to try something different.
GARCH.
In addition to MO one more real and widely used direction, but terribly labour-intensive and requiring a completely different training in contrast to MO.
Good luck.
With great pleasure I will see the results.
I've seen what you put into components, but I don't remember what you did with them.
GARCH.
Would love to see the results.
Yeah, I don't know how to do it myself, but I'm promoting it so that someone can figure it out for me.
I "know" a lot less than I "don't know". If you divide one by the other, it's infinitesimal.
Is it different for you?
GARCH.
In addition to MoD, another real and widely used direction, but terribly labour-intensive and requiring completely different training from MoD.
Good luck.
I will be very pleased to see the results.
Fourier is good because the cycles are well predicted on new data, but they do not reflect the complex dynamics of the instrument as well. Which is no longer the case with cycles found through spectral analysis. They are predicted much worse, but they reflect the dynamics better.
Based on banal logic, it is better to predict what is predicted better and then work with it somehow.
I didn't bother with a new TS, but added a filter on Fourier cycles to the existing algo. Roughly - the TS trades only when the cycle is in a certain range of values.
I have got something curious beforehand, but I need additional tests. Then I will finish the final one and test it in the terminal. There may be hidden peeking.
Small cycles as a filter:
Large cycles as a filter: already bad