Machine learning in trading: theory, models, practice and algo-trading - page 600
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
Look it up, but the problem is that there is not much fresh, systemic information about working with low-level NS structure at scale level, because our researchers rarely go down even to backend, like TensorFlow, mostly everyone is spinning above or at Theano, Keras, Torch or unfading R level.
No need to be silly. TensorFlow, Theano, Torch and CNTK are all low-level automatic differentiation libraries, used in deep neural network training. There are a lot of superstructures over them, one of the most common is Keras. For a regular user (not a neural network expert) it is convenient (easier and faster) to use high-level ones.
I am curious, what libraries did you use? Or just heard about them?
Good luck
Don't be silly. TensorFlow, Theano, Torch, and CNTK are all low-level automatic differentiation libraries used in deep neural network training. There are a lot of superstructures over them, one of the most common is Keras. For a regular user (not a neural network expert) it is convenient (easier and faster) to use high-level ones.
I am curious, what libraries did you use? Or just heard about them?
Good luck
Yes, you should read more carefully, not to talk nonsense.
It was about the back-end, I hope you don't need to explain the meaning and the front-end is for the user interface, a higher level.
And as for the interest, maybe you heard about the graphs TensorFlow, Protocol Buffers, code generation for different platforms and languages, i.e. essentially the low level, so I do the same, but for my NS and MQL language.
You probably haven't heard of it - Hlaiman EA Generator.
Forum on trading, automated trading systems and trading strategies testing.
Machine learning in trading: theory and practice (trading and not only)
Aleksey Terentev, 2018.01.23 06:39
Yes, I have difficulties. I have a hard time understanding why it's so hard for some people to get off their heels and work hard.
Yes, I hang out in this thread and interfere with other people's answers instead of constructively considering questions.
__Let's have a constructive discussion about deep learning? Using python? Learning with a teacher on good signals?
__I have no one to discuss with. And you just said you can't find anything. Well, that's not the case.
Yes, I'm giving a sucker's lecture. After all, my opponent himself used veiled mockery, confused terms, and made a couple of logical errors.
And I also apologized in advance, because I wrote in my feelings.
And also I offer you the help in knowledge of principles of work of tools for creation of neural networks. Without any irony and sarcasm.
I hope you understand from what I wrote above, the direction in which I may be interested in helping.
Visualization of graphs, NS topologies, serialization, ProtoBuf formats, batch processing and import/export of n-dimensional arrays NumPy weights of NS, etc..
If you have this kind of information or experience in implementation, with pleasure, ready to discuss.
Well, yes, and that not to say nonsense, you need to read more carefully.
It was about the back-end, I hope you do not need to explain the meaning, and the front-end is to work in the interface with the user i.e. a higher level.
Let's not argue about terminology. Here is an extract:
"Using the TensorFlow library".
Recently the rapidly growing area of deep neural networks has been extended with a number of open source libraries. Widely advertisedTensorFlow(Google),CNTK(Microsoft),Apache MXNet and many others. Due to the fact that all these and other major software developers are part of the R Consortium, all these libraries are provided with an API for R.
All of the above libraries are very low-level. For beginners to learn this area, they are difficult to master. With this in mind, the Rstudio team developed thekeras package for R.
Keras is a high-level neural network API, designed with a focus on being able to experiment quickly. The ability to go from idea to result with the least possible delay is key to doing good research. Keras has the following key features:
Friendly API that makes it easy to prototype deep learning models.
And about the interest, maybe you heard about the TensorFlow graphs, Protocol Buffers, code generation for different platforms and languages, i.e. essentially low level, so I do the same, only for my NS and MQL language.
Not only I've heard about it, but I use it as well. But with the R language for execution in MT. So we have a different approach and direction. My experience will not be useful to you.
You probably haven't heard - Hlaiman EA Generator.
I've heard of it, I've read it. Not the way I want to go.
Graph visualization, NS topologies, serialization, ProtoBuf formats, batch processing and import/export of n-dimensional arrays NumPy weights of NS, etc.
If you have this kind of information or experience in implementing it, I would be happy to discuss it with you.
I will repeat once again. We have a different approach and direction. My experience will not be useful to you.
Vladimir Perervenko:
For beginners to learn this area is difficult to master.With this in mind, the Rstudio team developed thekeraspackage for R.
Good luck
Interesting, I haven't seen a description of such tandems anywhere... I'll have to look it up
Back in 2007, they were building committees of 3-5 strategies and the quality of work improved greatly. But the problem in the committee is that at least two of the three must be adequate, then they will pull the committee to a greater advantage than separately. If the committee has 2 models overtrained. Trumpet business. At best, it will not let them merge, which in this case is not bad at all!!!!
Yuriy Asaulenko:
Vladimir Perervenko:
For beginners to learn this area is difficult to master.With this in mind, the Rstudio team developed thekeraspackage for R.
Good luck
I don't understand about Keras. Just yesterday I read that it's a high-level add-on to TensorFlow, and even saw an instance of it. No R, just Python.
The man even gave a link not to get lost, carefully avoiding R and even in these conditions managed.
I don't understand about Keras. Just yesterday I read, that it is high level addition to TensorFlow, and I even saw an example. No R, only Python.
What is there to understand. Everything in Python is already in R. Go to the links, take a look.
Good luck
The man even gave a link not to get lost, carefully avoiding R and even in these conditions managed.
Didn't see the link.
Here is the link to Keras -https://habrahabr.ru/company/ods/blog/325432/
I don't exclude that an interface to Keras is made for R. But it wasn't R that invented Keras. That is, it was not the Rstudio team who designed thekeras package for R, but rather the interface to Keras. And for the user, that's two big differences -- the package or the interface.
That's what I'm trying to clarify.
Didn't see the link.
Here's a link to Keras -https://habrahabr.ru/company/ods/blog/325432/
This is a link to the Hubr. The library link is https://keras.rstudio.com/index.html.
Read primary sources.
Good luck
Back in 2007 built committees of 3-5 strategies and the quality of work improved significantly. But the problem in the committee that at least two of the three must be adequate, then they will pull the committee to a greater advantage than the individual. If the committee has 2 models overtrained. Trumpet business. At best, it won't let them merge, which in this scenario is not even a bad thing!!!!
ensembles and committees are a little different than tandem, as far as I understand
by the way, the ensemble of the ns, the same MLP, is very good... but slow
about the committee is interesting, but controversial, the same ternary classifier Reshetov
I did not mess with tandems