Machine learning in trading: theory, models, practice and algo-trading - page 2489
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
And that makes you a neural network expert with 20 years of experience? It's so funny it's sad))
That I have VERY great experience in training and organization of AI systems, as well as a deep knowledge in the theory of NS. But you've apparently forgotten that I was a teacher and that I explain everything as little, so we have young people in the branch, so in fact I speak in simple language. But this does not make me a guru or an expert. I just share my experiences and observations that I have seen.
Just ask yourself what will be the apogee of ALL the work you do? Personally for me, it's a clear algorithm for getting models, no left or right to retreat to. It's the eternal jumping between methods, the eternal enumeration of data that you feed in or out. I do not have all this already, because I have found a way and use it all the time, and the one who is still looking, is doing what puts out variants of results and is always looking and looking and looking.... I have already found everything for myself, but I only add alternative data for analysis, so... so that it would not be boring!
That I have VERY extensive experience in teaching
as well as a profound knowledge of NS theory
Pushing buttons in 2 ready-made programs for 5 years does NOT make you an expert in networks with 20 years of experience, why this childish lie?
in training and in organizing AI systems
name a single author you've read about AI
I used to be a teacher and what I'm telling you is that I'm just a kid.
Writing posts on a forum is teaching?
Personally for me it's a clear algorithm of actions to get models, no left or right to deviate.
Do a test of your algorithm and that's it, it's better than 1000 words.
poking buttons in 2 ready-made programs for 5 years does NOT make you an expert in networks with 20 years of experience, why this childish lie?
So you do not understand the fundamentals of machine optimization important is not the model that you get, and that which you have chosen out of 10 equally obtained, because the algorithm uses random partitioning training, which leads to different end results, even if the metric values are identical, the coefficients found will still be different, and therefore will work differently. And since the question of trading models comes down to the choice of the obtained, the winner of this contest is the one who is engaged in selection for the practical use of models, rather than the one who is looking for new methods, combines old ones and digs in his own shit!
name a single author of a book about AI that you've read.
I haven't read any.
Writing posts on a forum is teaching?
It's more of an education, no more....
Do a test of your algorithm and that's it, it's better than 1000 words...
It's too time consuming to do a test every week and put them together, and the work from the real account I posted. and don't fuck with my brain!!!
p.s. Here's the deal
I checked everything again, found the error and built a few models again and everything seems to work. The pictures from left to right show the original dataset, the catbust forecast and the neural network forecast in 2 layers. The white areas in the model predictions show the areas with maximum probability <0.3
Thank you very much, everyone who responded and gave advice.
I checked everything again, found the error and built a few models again and everything seems to work. The pictures from left to right show the original dataset, the catbust forecast and the neural network forecast in 2 layers. The white areas in the model predictions show the zones with maximum probability <0.6
Thank you very much, everyone who responded and gave advice.
Forum on trading, automated trading systems and trading strategies testing
Discussion on article "Machine learning by Yandex (CatBoost) without learning Phyton and R"
Aleksey Vyazmikin, 2021.11.06 03:27
For those who don't understand some of the details on setting up and running the CatBoost described in the article, I decided to make a video clip by the example of EURUSD
The fish was found, but not much.
But purely out of the box, no tweaks - everything by the methodology. I did not expect to find anything at all, since the article is purely technical in methodology, and the predictors are mostly standard oscillators.
I didn't understand anything, but the music is still humming in my head)
Have you read the article? I can certainly add a test, but will lose the music :)
Write what exactly do not understand - I'll explain, just do not know what the emphasis there.