Machine learning in trading: theory, models, practice and algo-trading - page 330
![MQL5 - Language of trade strategies built-in the MetaTrader 5 client terminal](https://c.mql5.com/i/registerlandings/logo-2.png)
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
You just have to read a couple of books and look around on the Internet. I was in the same state a month ago). I started with SanSanych's article, and... I didn't understand a damn thing.)
Maxim Dmitrievsky has not quite a neural network yet, but the results are already good. And close to the middle of the topic someone published the results.
Nothing happens all at once).
Read about the neural network?
In that case, what should I understand?
Forum on trading, automated trading systems and strategy testing
Computer Learning: Theory and Practice (trading and not only)
SanSanych Fomenko, 2017.05.08 17:38
Why are you picking on the networks? They do not work and that's it, just a fashion of the past centuries, probably the first machine learning package that was available.
There are more promising ones: random forests, various ada. And in general caret shell package that has a couple hundred packages, including meshes, and you can do automatic selection between them.
PS.
Serious meshes that probably work here and here. The author is on the forum, cuts through the trade, links to MT4/5 terminals ...
PSPS
How can you seriously compare R to skylab? Some kind of rustic package, not in any rankings...
PSPSPS
And it's not about models at all, it's about datamining. You find predictors that relate to the target variable, you're golden.
Everything else is mind games.
Read about the neural network?
Then what is the meaning of this?
Neuronet forecasting is pure creativity, even if you thoroughly study them not the fact that you will find a model that will give you needed result... Neuronet is just a tool, you should at least decide what you want to analyze and why, what trading style, what TS is planned and so on... and only then you pick up a toolkit for your idea... so on... If you want to find TS with the help of NS, that's another question, then yes, data mining, analysis of predictors and so on...
If you wrote that you have a set of quotes, decide what to do with them... and then look for answers in articles :)
I don't know, if someone wanted to do some analysis with help of neuronet, but I was thinking that if you'd like to do some analysis with help of neuronet then you'd better think about what you'd like to do and why, what trading style, what TS is planned and so on... and only then you could choose a toolkit for your idea... If you want to find TS with the help of NS, that's another question, then yes, data mining, analysis of predictors, and so on...
The question is whether it is worth it to start and what is better.
To decide, I would first like to understand - what you can try to read to get such and such a result.
That is, if we approach the task from the end, I would like first of all to ask - does anyone have a noteworthy result of work?
No one shows anything...
Then what is the point of bothering?
For example, I showed mine.
I want to compare at the first stage.
If you get better results here, I would love to dive into this science.You wrote that you have a set of quotes, so decide what exactly you need to do with them... I don't know how to do it and then look for answers in your articles :)
I don't have a set of quotes, but a currency snake, which consists of a highly correlated set of currencies (not currency pairs).
I'm thinking, how do I do this properly?
The question is whether it is worth it to start and what is better.
To decide, I would first like to understand - what you can try to read to get such and such a result.
That is, if we approach the task from the end, I would like first of all to ask - does anyone have a worthwhile result of work?
No one is showing us anything...
Then what's the point of bothering?
Data Mining is not just neural networks and machine learning. This is a very large area, and the results of the real application of these technologies are amazing. It is amazing.
Among the problems solved, there are problems similar to those of trading, and successfully solved and solved by Data Mining.
The strategies that really work are becoming more and more complex every year. For example, I made my first strategy in 3 days on my own, and it worked for 2 years. Now it will not work with any settings - I tried it for fun).
The next strategy I have made for more than a month. About two years, and it has come to absolute dead state too.
I have been working on this last strategy for about half a year now. Trading logic is already too complicated. The only way out is Data Mining. SanSanych chose forests, I chose neural networks, who digs in other directions. In fact, there are no definitive solutions). Everything changes.
ZS "What's the meaning of this?" - SanSanych has his own opinion, it seems to him so. It is far from certain that he is right. But he also deals with other things - random forests, which is a whole other topic.
ZS2. For example, if those same neural networks recognize images (faces, for example), what prevents them from recognizing patterns for market entry. The task is more simple, by the way.
I don't have a set of quotes, but a currency snake that consists of a highly correlated set of currencies (not currency pairs).
I think - how can I trade it correctly?
I don't know a lot about it... If you could give me some more information about the snake, here are some links.
ZS2. For example, if the same neural networks recognize images (faces, for example), then what prevents them from recognizing patterns for entering the market. The task, by the way, is more simple.
Yes, but such a hat will need to be trained continuously for a month on a gtx 1080ti :) there are very complex recurrent nets that need a lot of examples for input so that they could wake up at least some recognizing intelligence :) Well this is what I know, computer vision takes a very long time to learn
I'm not too good at it... if you could give me some more information about this currency snake) some links
wiki, verbatim
Well, then the currency snake has long left its tunnel and no longer exists.
If you mean the cluster indicator (a la force of currencies), then I myself will be adding a grid to it soon :)
Yes, but such a hat will need to train continuously for a month on a gtx 1080ti :) there are very complex recurrent nets that need a lot of examples for them to get at least some recognizing intelligence :) well this is as far as i know, computer vision takes a very long time to train
Say, handwriting recognition is an uncomplicated neuron in several layers. There are even copies on the Internet, with training and a demonstration of the results. And one learns in front of an amazed public).
For market patterns, I suppose, if you work on 1m, then 2-3 weeks of history for learning is enough. In general, I'm going to start from the beginning.
In general, I'm going to solve a problem similar to your (Reshetov's) system, but on honest neuron, and see what comes out of it.