Machine learning in trading: theory, models, practice and algo-trading - page 3655

[Aleksey Vyazmikin]

Maxim Dmitrievsky #:

More often they make money from implementing and maintaining the ones they already have )

Well, like a business on, "We'll boost your sales with NS. We don't know how long the effect will last"?

[Maxim Dmitrievsky]

Aleksey Vyazmikin #:

You know, like, "We're going to boost your sales with NS. We don't know how long the effect will last"?

I don't know about that... but go to any supermarket and try to find a salesman in it :) cash registers are getting smaller and smaller. Savings? Savings.

[Aleksey Vyazmikin]

Maxim Dmitrievsky #:

I don't know about that... but go to any supermarket and try to find a salesman in it :) cash registers are getting smaller and smaller. Economy? Economy.

And I only use cashiers, I'm kind of annoyed that I'm asked to do the cashier's work myself.... if there was a payment for it in the form of a discount, I would think, but so... I'm taking bread from a lot of people.

[Maxim Dmitrievsky]

Aleksey Vyazmikin #:

And I only use cashiers, I'm kind of annoyed that I'm asked to do the cashier's work myself.... if I was paid for it in the form of a discount, I'd think about it, but this way... I'm taking bread from a lot of people.

It's everyone's business, some people eat fruit and sweets that are by weight and walk away ))

[Aleksey Vyazmikin]

Maxim Dmitrievsky #:
Everyone's business, some people eat fruit and sweets that are weight gainers and go away ))

That's just the way it is.

[Alexandr Sokolov]

Maxim Dmitrievsky #:

... I'm 98 out of 7,500, and I'm still hanging on.

Congratulations! - That's a decent result :)

[Maxim Dmitrievsky]

Alexandr Sokolov #:

Congratulations! - decent result :)

Thanks, but I didn't try hard enough. Now I have to poke around to find out their methods. They said they will publish the approaches later :)

And it should be very strong, if not a random feature brute force (often those win). But the task is pretty hard to hit by chance.

[Maxim Dmitrievsky]

A fast tester on Numba now allows you to fiddle with TCs on minutes and even ticks, given the ability to accelerate other phs than MO algorithms, which are already mostly in C.

In pure python this was a luxury.

[Maxim Dmitrievsky]

The slowest thing is learning. It makes sense to consider either spike NSs as the fastest (if you take NSs) or bousting over linear models. I'll have to dig into the topic.

[СанСаныч Фоменко]

Does anyone know what the r-bloggers.com site is doing? I haven't had it load for many days.