Machine learning in trading: theory, models, practice and algo-trading - page 2769
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
Signs will be the price spread over the entire history of the instrument divided by the price step)))) And their combination. It seems not the purpose of the number of signs. There are curses.))))
Zy, only without TRIZ))))).Well, that's right, it's a pity that there are no substantive discussions of these steps, although everyone here has them, but apparently for some reason they are considered secret))))
There is nothing wrong with TRIZ, as long as it is used as intended))))) The paradigm of correct decisions from the correctness of the approach is generally obvious))))))
I am trying to describe extremums, I came to the conclusion that it is necessary to describe not the extremums themselves, but what is inside / between them, then you can get some description of the state. Also patterns are mostly too different inside to be meaningfully informative. Machine sifting of patterns is certainly better than visual sifting, but it is quite possible to look inside on the lower TFs.
Just some small thoughts)))
Well, it's true, it's a pity that there are no substantial discussions of these steps, although everyone here has them, but apparently for some reason they are considered secret))))))
There is nothing wrong with trise, as long as it is used as intended))))) The paradigm of correct decisions from the correctness of the approach is generally obvious)))))
I am trying to describe extremums, I came to the conclusion that it is necessary to describe not the extremums themselves, but what is inside / between them, then you can get some description of the state. Also patterns are mostly too different inside to be meaningfully informative. Machine sifting of patterns is certainly better than visual sifting, but it is quite possible that we need to look inside on the lower TFs.
So, small thoughts))))
If there are more patterns than bars in the history, there is something wrong with them already
I hope for fractality) And maybe some dynamic descriptions. Dynamic characteristics usually cover whole groups of static ones.
The task of describing the state of a process close to the SB, it doesn't matter why it is close, whether it is the SB or the sum of functions not known to us, it doesn't matter, the task is close in essence to the search for standard states in environments similar to the SB)))))
But the goals are different and accordingly the approaches of solutions.
I like this task more than searching for something in chaotic environments.
I hope for fractality) And maybe some dynamic descriptions. Dynamic characteristics usually cover whole groups of static ones.
The task of describing the state of a process close to the SB, it makes no difference why it is close, whether it is the SB or the sum of functions unknown to us, it makes no difference, the task is close in essence to the search for standard states in environments similar to SB))))
But the goals are different and accordingly approaches of solutions.
I like this task better than searching for something in chaotic environments.
You don't take into account the market structure itself, you look at the abstraction... SB or sum of functions....
Of course I don't take into account, and I don't set such a task, I'mafraid) )))) I can't cope))))))
I set myself a simpler task, to describe some stable states (not all of course)))) ) through some indicators obtained from history. Weather is described not only by temperature, but also by pressure, wind speed, humidity, sky transparency. We initially have only 2 parameters, price and time. They are definitely not enough to describe the state.
Of course I don't, and I don't set such a task, I'mafraid as it is) )))))) I can't cope))))))))
I set a simpler task for myself, to describe some stable states (not all of them, of course)))) ) through some indicators obtained from history. Weather is described not only by temperature, but also by pressure, wind speed, humidity, sky transparency. We initially have only 2 parameters, price and time. They're definitely not enough to describe the condition.
An unprecedented thing happened - in the MO branch they accidentally noticed that quotes are not an abstract series of double values :-) but they try hard to ignore it, because it's difficult
trading EURUSD (euro vs dollar). what do we see?
at least 4 categorically different states: (1) It's day in Europe and national banks change currencies (and the exchange back and forth to the dollar, this is Europe, the states don't need it, they have enough dollars), (2) the states are switched on and the revaluation of funds starts (fixed assets in the states), (3) Europe ends, the states remain, (4) only all others are left.
this is superimposed on the structure of the session: three pronounced peaks in volume - a clear substantial one at the opening, a slightly smeared one after the local lunch and a small final one before the closing. That's how banks work.
And to keep things interesting, leading exchanges organise fixing - at a strictly scheduled time they evaluate potential rates (and conduct corresponding trades). This will be the reference price, but the "big companies" will try to trade actively at this time (this is their "arrow" - it is convenient for them to trade with each other). At these happy times, there are NO TICKS.
That is, there is already a hell of a difference :-)
And as a cherry on the cake - our distributed forex, where each node can be represented as a mass service system; which again has a couple of states - it has time to service orders and react to changes in its neighbours (ticks are regular, 1-2 points each) and it does not have time (ticks are at a "jagged pace" and changes are significant)
Well, I've achieved on cloze alone what bousting with a hundred features can't achieve....