Machine learning in trading: theory, models, practice and algo-trading - page 1654
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
Finally I made a crumpled output to the TA program from my AMO.
I say right away it's just a draft, which is deprived of much, such as adaptive adjustment, which should be on every bar (I have not yet implemented it in the output), that is, the algorithm trades yesterday's morning knowledge of the market, and this is certainly not optimal, but we will finish it.
The algorithm should also be considered as the first block of three:
block 1 : market direction
block 2 : best entry point
Block 3 : filtering false entries.
That is the entry points are not clear not nice, but it is not the task of this block, the task of the block to be adequate to the direction of the market
Trading is essentially without parameters, one logical rule for entering a position, and that's it.
This is the Euro 5 min. if anything) (the most unpredictable and everyone's favorite instrument)
I wonder what will happen to him by the end of the day, because he's already using not actual knowledge about the market ... Let's see...
It's Friday, the market should be sharply moving, it will be a stress test.
All in all, it looks interesting. The main thing is not to cheat yourself. The main thing is not to deceive yourself by believing in nonsense not related to the subject area. In contrast to another opponent, who, although making predictions on the bitcoin, but how to use it he does not even know. In your case, everything is clear and local. The arrow up to buy, down to sell. No questions.
All in all, it looks interesting. You see what happens when you have an adequate approach to the matter, not flying in the clouds. The main thing is not to deceive yourself by believing in nonsense not related to the subject area. In contrast to another opponent, who, although making predictions on the bitcoin, but how to use it he does not even know. In your case, everything is clear and local. The arrow up to buy, down to sell. No questions.
Right, right, right.
.
Yeah yeah yeah
Again I do not know the essence of your TS, but I still wonder why nobody uses VTRIT for P, which sifts out inputs and makes their preliminary culling. There's an option for predictive regression in addition to that for classification. But for that you need to have a set of inputs to cull. For example at me from 7000 inputs only 150 or so remain after rejecting P. And resulting model never contains more than 11 inputs. Such a weird set, even though there are only 10 indicators at the base.... Surprising to be honest...
I don't know what you're talking about.
Well R has an addon called vtreat. So it's like a really cool way to improve the system's performance. In general, about him. But only if you have a lot of input variables and you want to reduce them, leaving only necessary. So.... By the way...
There's a lot in R
P.1.2 is interesting, trend-oos.
didn't get it)
didn't get it)
In a nutshell:
Study period.
OOS or test period.
Entry conditions.... So that you know what we're talking about, maybe we'll tell you something valuable... Anyway, I'm a teacher... :-)
In a nutshell:
Study Period.
OOS or test period.
Input conditions.... So that you know what we're talking about, maybe we'll tell you something valuable... Well, me anyway, I 'm a teacher... :-)
My teacher -
The forgotten past - the arrows...
Enlighten me...