Machine learning in trading: theory, models, practice and algo-trading - page 2157

[iwelimorn]

Maxim Dmitrievsky:

I need an efficient function that will put the spread in the feature space

you can try it this way

prices_l = prices.get('low').reset_index(drop=True)
prices_h = prices.get('high').reset_index(drop=True)
sp = prices.get('spread').reset_index(drop=True)/100000
 close_diff = prices.get('close').reset_index(drop=True).diff().mul(((prices_cl- prices_op)-sp)/(prices_cl- prices_op))

then on the basis of close_diff calculate the target, or sample a random trade.

I'm just sawing the target and it looks like this

[Uladzimir Izerski]

Aleksey Vyazmikin:

Why give an assessment to the "Forumers," what is the purpose of this task?

I am impressed by your ability to achieve your goal. Such people can always find their place in a team. The collective nature of business prevails over the individual.

It is possible to create a working group for a serious task.

[Maxim Dmitrievsky]

welimorn:

you can try it this way

then based on close_diff calculate target, or sample a random trade.

I just saw the target and got it like this

Cool! I will try it, thanks.

That's what I wanted to do.

[iwelimorn]

Maxim Dmitrievsky:

Cool! I'll give it a try, thanks.

That's what I wanted to do.

above error, we need to add the spread:

 close_diff = prices.get('close').reset_index(drop=True).diff().mul(((prices_cl- prices_op)+sp)/(prices_cl- prices_op))

and now it looks like both classes are bunched up))

[Maxim Dmitrievsky]

welimorn:

above error, we should add the spread:

and now both classes seem to be bunched up))

I'm almost asleep now... I'll see later. Well, if it doesn't work out, there's another complicated option, I'll describe it later, maybe if I get to

[Valeriy Yastremskiy]

Maxim Dmitrievsky:

You have to work with spreading densities for each group of labels. I don't see any other options yet, only hardcore variation

with changing densities and finding points between

acceleration diffs... Just by hand the accelerations are not significant

[Valeriy Yastremskiy]

Maxim Dmitrievsky:
I'm almost asleep... I'll see later. Well, if it does not work - there is another complicated variant, I will describe it later, if I get there

you have to sleep at night)))) with this time zone))))

[Aleksey Vyazmikin]

Uladzimir Izerski:

I am impressed by your ability to achieve your goals. Such people can always find their place in a team. The collective nature of business prevails over the individual.

There is an opportunity to create a working group for a serious task.

Thank you for the flattering assessment.

You're a master at evading answers.

So what do you have to offer?

[Aleksey Vyazmikin]

Valeriy Yastremskiy:

I went to first grade in 73, and he was in the army.... consider... fathers and children the eternal book)))))

Do you think it's an age thing?

I've noticed that nowadays pensioners don't sit near the porch and don't gossip as they used to...

[Maxim Dmitrievsky]

Aleksey Vyazmikin:

Thank you for the flattering assessment.

You're a master at evading answers.

So what do you want to suggest?

Apparently the same as the Latvian.