Machine learning in trading: theory, models, practice and algo-trading - page 563
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
Whew... read the whole thread from the beginning to my appearance... now I've seen everything
but did not find the grail there ... too bad, I'll keep digging mine then
This is the right solution. NS theory here on the forum is far from ideal.
The only thing I wrote down was about the ternary classifier, and solving the mystery ofYuriy Asaulenko
This is the right decision. The theory of NS here on the forum is far from ideal.
The only thing I wrote down was about the ternary classifier and solving the mystery ofYuriy Asaulenko
The only thing I wrote down was about the ternary classifier, and to solve the mystery ofYuriy Asaulenko
And where did you find the mystery.
MLP is ~60 neurons. The algorithm is standard BP. Learning - go hither and thither. i.e. I don't know what the NS is learning there. In addition, all the principles of learning are outlined in classical monographs - Heikin, Bishop. Soft - not MQL.
The basic principles are, in my opinion, described in this thread.
And where did you find the mystery.
MLP is ~60 neurons. Algorithm - standard BP. Learning - go where I don't know where. i.e. I don't know what the NS learns there. In addition, all the principles of learning are outlined in classical monographs - Heikin, Bishop. Soft - not MQL.
The basic principles are outlined in this topic.
This was kind of a joke :))
that was kind of a joke :))
No. There's really nothing else there. You think that Haykin and Bishop are hopelessly outdated and search for something new. They're good enough for me.
No, I mean it's like I was joking... you're the only one in the thread who came up with something in the end :)
you need to google perceptron training by the monte carlo method.
In general, this method is very similar to RL (reinforcement learning) when there is a learning agent and the NS is learning to find the best solution
This is how Alpha Go is trained (although it was previously assumed that it was a creative game and a machine could not beat a human in it)
And here's the winner.
https://techfusion.ru/nejroset-alphago-pereveli-na-samoobuchenie/
No, I mean it's like I was joking... you're the only one in the thread who came up with something in the end :)
you need to google perceptron training by the monte carlo method.
In general, this method is very similar to RL (reinforcement learning) when there is a learning agent and the NS is learning to find the best solution
By the way, it's largely thanks to you. When I just started, it was you who gave me the link to Reshetov's article. The article, in general, is nothing, rather as an example of application, but it became approximately clear where to harness the horse.
I do not know, whether there are such methods in Google, as I myself have finally come to Monte Carlo.
I don't know about RL either, but from your brief description it sounds like my methods.
I found Monte Carlo in Google -https://logic.pdmi.ras.ru/~sergey/teaching/mlbayes/08-neural.pdf Only this is completely different.