Machine learning in trading: theory, models, practice and algo-trading - page 2339

[Sergey Chalyshev]

It might help

[Evgeny Dyuka]

We've been working too long...) We can't finish it, and we won't make any money. Something is getting in the way.

It's ok, I'm testing it with real market, it works fine.

[Aleksey Nikolayev]

Mikhail Mishanin:

Pandemonium of reasoning/thinking https://ai-news.ru/2021/02/prostye_i_maloizvestnye_sposoby_veroyatnostnyh_ocenok.html

However, the wizardry of the great Bayesian estimation of the binomial distribution parameter (apparently, using the beta distribution as its a priori distribution)

[mytarmailS]

BillionerClub:

The super-duper smart machine seems to be discussed in the thread. I have a question, as a matter of fact, what prevents machine learning to make money on forex, just as the AI plays Dota with the champions and defeats the latter?

If you had asked how much it costs to train such a model, how many experts develop it... maybe you wouldn't ask such silly questions...

[Maxim Dmitrievsky]

If you teach NS to trade the grid\martin, you get a similar picture, if trained without it. Still the same pattern disappears in 2015-17

probably already squeezed the maximum out of the hour bars, it goes no further

a couple of examples

[Forester]

Maxim Dmitrievsky:

If you teach NS to trade the grid\martin, you get a similar picture, if trained without it. Still the same pattern disappears in 2015-17

probably already squeezed the maximum out of the hour bars, it goes no further

Some examples

Regular martins tend to fail at the end of the day. Does an NS martin also lose?

[Maxim Dmitrievsky]

elibrarius:
Ordinary martins tend to fail in the end. Does an NS martin also leak?

how you set it up.

[[Deleted]]

23,000 posts!

Guys, maybe it's time to come to some kind of conclusion!

[Konstantin Nikitin]

Vladimir Baskakov:
23000 posts!

Guys, maybe it's time to draw some conclusions!

Everyone draws his own conclusions. Someone completely relies on the NS. And someone uses it as an additional filter for their strategy. There are those who do not trust the NS at all.
This topic is more like a demonstration/assessment of possibilities and drawbacks I've noticed...

[Maxim Dmitrievsky]

For those who don't know what to do

https://stefan-jansen.github.io/machine-learning-for-trading/