Machine learning in trading: theory, models, practice and algo-trading - page 1184
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
art. )
But in general, this particular problem is solved by crossvalidation. Another thing is that in most cases with cv it will turn out that laws describing the general population simply do not exist in the training sample
art... )
But in general, this particular problem is solved by crossvalidation. Another thing is that in most cases cv will turn out that laws describing the general population are simply not present in the training sample
it's not so simple, I've been trying to visualize the MO since a week, and just in the tester to fit the story is possible without the MO, here look how you need to prepare the workplace for the MO.... I'm going to take Windows down on my laptop... Maybe I should go for Peony...
art... )
But in general, this particular problem is solved by crossvalidation. Another thing is that in most cases cv will turn out that the laws describing the general population simply do not exist in the training sample.
The point is that we must be sure at least in the presence of similar events, then you can look for similar ways to describe them, at the expense of CV.
I understand that neural networks will not work in the market, at least as it is without transformations.
The NSs will work in the market as they are, without any transformations. Transformations are needed for the market data to be fed to the NS inputs.
The NSs will work in the market as they are, without any transformations. The market data need transformations in order to feed them to the NS inputs.
Are there any working examples? At least to look at this miracle?)
Well, the whole topic is devoted to it. There are examples of such transformations on almost every page).
I understand that neural networks will not work in the market, at least not as it is without transformation.
Well how not, in a dozen lines you can make a neural network-based MA and such a neuroMashka will work exactly the same as a conventional MA, NS do not give a miracle, but simplify data analysis and nothing more, the NA is not able to think, the one who solves the problem with the help of MO should think
Do you have any working examples? At least look at this miracle).
If you are a "programmer from God" then I would recommend you not to do what everyone else is doing - those trivial things - those are to look for patterns in the price using the price itself.
I can give you some tips, but I don't think you'll do it, because non-standard way of thinking scares you away.
If you are a "programmer from God" then I would recommend you not to do what everyone else is doing - those trivial things - those are to look for patterns in the price using the price itself.
Something I can tell you, but I do not think that you will take it, because non-standard thinking scares away
Well, at least tell us. Briefly.
I wonder what to do, what everyone else is not doing. It seems to be all options.
I promise not to laugh out loud.