Machine learning in trading: theory, models, practice and algo-trading - page 2245
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
It all depends on the task, I do not understand what you want to hear
I want to see an effective application of DSP to quotes from those who have been doing it for 100 years
I can take these filters and give them ready TS in the form of a bot, with history tests, etc. But they don't want to. But they don't want to. Or rather, they have nothing.
There's nothing complicated about replacing some MAs with others
I want to see an effective application of DSP to quotes from those who have been doing it for 100 years
Well, it's definitely not up to me
I can take these filters and send them ready-made TS in a bot form, with history tests and so on.
Well you can synthesize these filters yourself using neuronics, you do it all the time ...
At input you have signs with noise and output - buy/sell.
A filter? Filter!
He wants everything to be chewed up and put in his mouth, and all he has to do is eat it.
It is not very clear to me, either. Of course, it is possible to call a filter to highlight signals to buy or a trend, and its completion. But in general, a filter must sift out what is necessary. The conditions of necessity may be anything, such as sea breakers, conditions of employment by gender), frequencies, conditions of characteristics repetition.
In our case, we have a series. We can decompose it. Decomposed, we have allocated what is considered significant. Then how?
For (meaningful application of) filtering, the signal must be a mixture of some useful raw signal and random noise. Like the above-mentioned sine wave plus noise, or aircraft trajectory with errors of its measurement in radar, etc. It is quite possible when the initial signal is not deterministic, but also random, but in its characteristics very different from the noise superimposed on it (for example smoother) - I saw an article where the spread was modeled as an Ornstein-Uhlenbeck process with added white noise.
The problem here is that having a priori and reliable knowledge about the decomposition of price into noise and some original signal is either a Grail claim or an element of schizophrenia.
I've shown you before.
Here's more for clarity and understanding:
code mashkas in the studio, or show it somewhere else
The problem here is that the mere existence of a priori and reliable knowledge about the decomposition of price into noise and some original signal is either a Grail claim or an element of schizophrenia.
I simply explained why a hundred trees is better than one...
I didn't just say 100 is better, I explained why.
That's all, no other purpose...
If you have no desire to describe all the nuances in detail, then deign to show an example of your trade - publish a paid signal, at least for one year.
All will appreciate your skills in the market, and at the same time, for this time and a million bucks from subscribers. What are you waiting for?)
I've shown you before.
Here's more for clarity and understanding:
.
I've been practicing the MA's myself for a long time. They are too late or too noisy. And the price at the moment of the signal may be in any place of the candle, and not necessarily in the right one. If you want to understand the logic in details.
Here is more for clarity and understanding
Thesystem of trend following, deals are opened when the filter has one direction on 4 TFs at once?
code of the mashcats in the studio, or show it somewhere else
there is no studio here; these are not mashups as you understand them; the code is my development and is not distributed; this thread is not yours, you are a guest here, and your rights are guest.
I have already shown too much, it's enough.