Machine learning in trading: theory, models, practice and algo-trading - page 1702
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
If it had been about regression, it would have been clearer - it probably should be taken that way.
I agree. NS is more of a component of AI, like a screw or a nut in a car.
Another one of your mistakes as a beginner. NS is very demanding to the training sample and is quite a delicate tool, where a small error in data preparation (a comma in the wrong place) leads to a diametrical result. Try to do less of your own conclusions, and better listen to what you are told.
All right, the NS is a component of AI. In addition to the neural network, there are a number of algorithms that serve this neural network, and together this is the AI system. Exactly a system. But an AI system cannot exist without a neural network.
At the moment I don't identify AI and NS. You yourself said that you shouldn't confuse the two. NS is a tool that can be used in AI, but on its own, it does not reach it. You seemed to agree with that.
Is the AI even required to interact with humans ???
So that's the conclusion we've come to.
What's wrong with it? Catbust bypasses most of its competitors.
For example I don't like the idea of a symmetric tree. Obviously it's not the best solution to divide 2 different nodes by one predictor and by the same level. Unless it speeds up to 10 times.
Good thing they added 2 new more classic methods.
Probably bypasses, but there, in contests, the sampling is stationary, there are not particularly trashy features, i.e. the conditions are not those with which we work, and I just think how best to prepare data with these features in mind. (No solution yet in final form, but it's an important task).
The different tree models are good, but so far it is not possible to upload them to a separate file, and thus it is not possible to embed them in the Expert Advisor, which is bad.
I don't like the absence of postprocessing in boosting - when at the end of training the model is simplified by throwing out weak trees. I don't understand why they don't do this.
Leaves of individual trees in boosting are weak - small completeness - less than 1% and it is bad that this parameter cannot be adjusted, and accounting signal distribution by sample is not carried out at all - as a result we learn on discarding. A lot of nuances and here the solution may be a good pre-processing of predictors. Better yet, of course, plug in and flesh out the code - no one really understands C++ at the proper level?
I think he has to. Otherwise, what's the point?
when you think (your intellect solves a problem) do you have to communicate with someone at that moment?
You still can't get your "worm" definition of intelligence out of the way, so we're communicating in different languages now
I'll tell you even more, I had such cases when I was left to press the optimization button and then I realized that here was the step 20 operations back I did wrong and in fact I had an error in the data and had to prepare all over again because I clearly understood that there should be no error. Literally a comma in the wrong place and all for nothing. And that is machine hours and time, and most importantly the result.