Machine learning in trading: theory, models, practice and algo-trading - page 1368
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
They promise to expand the functionality over time. If suddenly, for example, it will be possible to manage testing/optimization from R, then it makes sense to switch.
If they expand it, then we'll see, now I don't see any problem to get quotes in another way
I don't need much, only the possibility to use cool ML libs like TensorFlow
The rules of pronunciation are not very important here-it was one of the ways in which Soviet humanitarians were hierarchical.
Good Alexstein Nikenstein.
I admit that it was a mistake to develop psychological topics in this thread, but they arose as a consequence of reflections on the human brain, the nature of its mind, and artificial intelligence, which seem to be hidden in the context of the topics discussed here.
I admit that it was a mistake to develop psychological topics in this thread, but it arose as a consequence of thinking about the human brain, the nature of its mind, and artificial intelligence, which is sort of hidden in the context of the topics discussed here.
Respect. Intelligence is not psychology, it is a separate and unrelated phenomenon.
In previous posts, the neural network predicted the shift of the assumed center of the assumed distribution of the assumed price. The prediction interval was 5 m. And everything looked good compared to the actual shift.
Now I decided to monetize the benefits of the prediction results, and see if I could actually benefit from this prediction. To do that, I compared the results of forecast and real price shift in the 5 meter interval on the chart.
Intelligence is the brain's ability to feel with its oops..... IMHO!!!
In general I've changed the Reshetov optimizer to my needs. Basically now focused on the selection of the resulting models, which in the process of running it turns out quite a lot. How to choose the one that will work in the future. Because the most optimized model is not always adequate to the market. In the control section, the logistic evaluation function, made him save the intermediate models during the optimization, not to mention the fact that he saves the code for my needs. I can't say that I've become a Java guru. But I'm already confident in editing other people's code... Still, I'm interested in the question of how someone selects models? What are the criteria used to decide that one model is better than another? I have a way without conditionally and it is quite interesting. In any case, he sifts out exactly the bad ones. But among the good ones we can find neither one nor the other.
I'll tell you what. Throw here the answers to my questions. Then I will tell you how and what I do..... Sharing experience I suggest to start with you :-)
That's how your intellect is built, or rather your butt instead of your brain
Well, from the heart. Do you have anything to say on the subject?
Well, from the heart. Do you have anything to say on the subject?
No, you did not say anything, I'm your information bureau or something.
Here I have nothing to say, as no one will need it. Interesting? Yes, but it is unlikely anyone will apply in practice, they will say blabbermouth Ha Ha. And it looks like you are the only one left here Maxim, scaring everyone away with your single all-knowing presence. This is one of the fundamental features of the field of machine learning. There are only three of them. Three postulates that I wanted to talk about in my video, which will surely come someday. There I tell you why the advice you give is interesting, but no one will use it. Rarely will they listen, but that's all. It's the problem of MO that the researcher fails.
This is the Achilles' heel of Machine Learning.