Machine learning in trading: theory, models, practice and algo-trading - page 339
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
Why is a sigmoid used to calculate a neuron?
Why is a sigmoid used to calculate a neuron? Isn't it better to have a linear distribution (from zero to the number of inputs)? After all, "the function has a smooth form on the interval [-5,5]".
It's good if you have only 5 inputs, but what if you have a hundred? Then practically all of the inputs will be outside of that area. The article https://www.mql5.com/ru/articles/497 applies an addition factor to account for 10 inputs. So for each network, you would have to recalculate this factor.
I wonder, didn't you find more recent articles on neural networks on this site?
I have the impression that I got to a forum ten years ago. Surprised
I wonder, have you found any more recent articles on neural networks on this site?
I have the impression that I got to a forum ten years ago. Surprised
I liked https://www.mql5.com/ru/articles/2279 using ALGLIB, but firstly it is not adjusted to the trading Expert Advisor and needs more work,
Secondly, as he says.
Maxim Dmitrievsky:
I'm not going to use MLP for market forecasting, see above videos, I need RNN, i.e. network with memory
Also I would like to have an implementation without external program, but purely in MT5, and to be able to distribute calculations across cores and network with cloud - for speed. Or can external network programs do this?
here is the link
what is interesting I have never seen free libraries with similar implementation of genetic programming...all only networks networks networks....
Genetic programming package https://cran.r-project.org/web/packages/rgp/index.html
Good luck
I liked https://www.mql5.com/ru/articles/2279 using ALGLIB, but firstly it is not finalized as a trading Expert Advisor and needs further work,
If I missed such a network, with an example of implementation in MT5 - advise)And secondly, as elibrarius says
I would also like to have an implementation without external program, but purely in MT5, and to be able to distribute calculations across cores of the computer and into a network with the cloud - for speed. Or external network programs can do it?
Take a look at This, This, This, and maybe This.
Not everything will be clear, but some basic neural network concepts will hopefully come through.
Good luck
Take a look at This, This, This, and maybe This.
Not everything will be clear, but some basic neural network concepts will hopefully come through.
Good luck
Take a look at This, This, This, and maybe This.
Not everything will be clear, but some basic neural network concepts will hopefully come through.
Good luck
I'm studying R as well :) Do you know where I can find LSTM, maybe some other links on the subject?
I also have started studying R :) Could you please advise where to find LSTM, any other links on this topic?
R has a nice mxnet package. But more advanced models should be looked at in Pythone.
Good luck
R has a nice mxnet package. But more advanced models should be looked at in Pythone.
Good luck
Thanks, should there be some in tensorflow?
Thank you, should tensorflow have it?
I don't think you can start with TensorFlow. This is a low-level package. There is a sea of packages in Python that implement different kinds of LSTM. But that's the hard way.
If you program in Python start with keras, R has a package called kerasr (API).
For TensorFlow(Python) there is also an API in R - tensorflow. Try it out.
Good luck