Machine learning in trading: theory, models, practice and algo-trading - page 3106
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 say here that the t-test sometimes works well on abnormal data as well
Yes....
But we still have a few years, or months, to go))
So far, there are two problems for launching a strong AI
1. Too voracious architectures
2. Too weak hardware
It's basically two sides of the same coin....
But work is underway to solve both the first problem and the second....
They are not in a hurry to change the architecture (neural networks are all we have), but theywill have to, but they are much more active with fast hardware (quantum computers).
There is a threat from AI, but in a vague distance and only after we answer the question: what is natural intelligence? As of today, there are no approaches to it.
It was a long time ago, but it has nothing to do with stationarity. Less than one indicates the stability of the model, it indicates that: correctly (1) differentiated, (2) modelled the variance (lots of models), (3) modelled the mean (ARIMA-AFRIMA), (4) modelled the distribution. In short, they are trying to model NOT stationarity.
Can I see the actual code of application of these garches? Or is it just a paraphrase of brochures without a drop of practice?
Can we see the actual code for the application of these garças? Or is it just a paraphrase of brochures without a drop of practice.
Tried a few years ago (2017-2918) but gave up - too complicated. Evaluate rugarch:ugarchspec. To add to that, the parameters are interconnected, it's all tied up in optimisation, a step to the side and you get hours of model fitting. I wasn't impressed with the results, but that's my fault, not because of the curvature of the model.
1) Tried it a few years ago (2017-2918) but spit it out - too complicated.
2) Not impressed with the results, but that's my fault, not because of the curvature of the model.
So why do you advertise this rubbish here regularly????
I don't want to try anything, I have already tried it for many years to come....
I can tell without trying what can work and what can't....
If the algorithm looks at the market as a time series, then goodbye at once, no matter if it is a stochastic or a praised Garch.
The result for me is already predetermined
This is the second or third time I've come to this thread to take a look. Nothing has changed, only thousands of pages have been added. Whether it's the first or the last one. it's the same.
What's supposed to change, machine learning only works on static.
to predict the future, it's nonsense.
What is the best clustering method for grouping such objects?
Basically there is a matrix, and it is important to evaluate its similarity as a whole. And for some reason, K-means, I think, will average everything out a lot.