Machine learning in trading: theory, models, practice and algo-trading - page 714
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
will be
Is there any way to feed this MA into a neural network for analysis, or is there no point?
Once again, I say that everyone has his own Grail, hard-won and coveted. If you've found it, use it. Why stick it where it's not necessary?
There's only one grailer here - me.
Once again, I say that everyone has his own Grail, hard-won and coveted. If you found it, use it. Why stick it where it's not needed?
There's only one grail-giver here - me.
Alexander, everyone has a grail in the terminal, and I showed it to him.
The only question is who and how much can squeeze out of it.
Intensity of tick quotes for AUDCAD pair (right chart)
Observation window = 8 hours, reading frequency = 2 sec.
No neuronet will ever make predictions until we find a Person who knows how to work with the intensity.
The author of the branch in his blog described an experiment where he predicted 18 points ahead. And for each he made a separate forecast by a separate MO system (forest, I think, from gbm).
Isn't it better to make one system (forest/NS) to forecast all outputs at once?
By the way, where did he disappear to?I understand that having 18 outputs you should have a lot of neurons in hidden layers and calculation will be long. But to calculate 18 separate systems is probably even longer?
The author of the branch in his blog described an experiment where he predicted 18 points ahead. And for each he made a separate forecast by a separate MO system (forests, I think, from gbm).
Isn't it better to forecast all outputs by one system (forests/NS) at once?
By the way, where did he disappear to?I understand that having 18 outputs, you have to have a bunch of neurons in hidden layers and the calculation will be long. But to calculate 18 individual systems is probably even longer?
I've seen his live monitoring somewhere, with low yields, but it seems to work.
in general, nothing very interesting.
He has a live monitoring there somewhere I saw, with a low yield, but it seems to be working
Nothing interesting in general.
The question was - "Isn't it better to predict all outputs by one system (forest/NS) at once?"
And in general, what are the pros and cons of calculating N outputs by one system and N systems with 1 output each.
Forecasting and classification is not trading. Even if you get satisfactory training data, it's not so easy to make it all trade. This is the difference between theory and practice...
As I and a couple of other participants repeated it more than once, but everyone's attention is focused on the next typical library (package, parameter configuration), as before on another magic-industrial in JMA style and so on.
Those whose hands don't grow out of their... those whose hands don't grow straight away have already understood that one can't get prediction accuracy higher than 55% for 1 min, while it's normally 52-53% and the correlation with (Close-Open) the next candle is about 0.05 (R^2 = 0.0025), moreover this forecast is very noisy, while averaging takes all the advantages, but this is the reality, the truth which has to be adjusted to. I personally don't know how yet((( no put strategy comes out.
Well the blog is very interesting...
The question was "Isn't it better to predict all outputs at once with one system (forest/NS)?"
And in general, what are the pros and cons of calculating N outputs by one system and N systems with 1 output each.
Well, in theory it makes no sense, because the NS should work well in multidimensional space and divide into any number of classes
it is better not to divide anything into classes and not to put labels
this way the learning process will be more correct, but more difficult to implement
q-learning rules
https://www.youtube.com/watch?time_continue=1685&v=ZkZQwKizgLM
I can download more training videos and examples in python for those interested