Machine learning in trading: theory, models, practice and algo-trading - page 2050

[Valeriy Yastremskiy]

    txc = GetTickCount();
         if(txc-tx<500)                    // в миллисекундах меряет, в той версии в секундах мерял.

mytarmailS

[[Deleted]]

dr.mr.mom:

lines:

1. 3+1=4

2. 2+1=3

3. 3+flag2=5

columns:

1. 3+2=5

2. 1+3=4

3. 1+flag2=3

with maximal non-repetition, purely imho

also true, but here without taking into account the convexity/convexity of the figures

it's a double situation

[Rorschach]

Maxim Dmitrievsky:

Who's going to figure it out? )

Raven's test? 5

[[Deleted]]

Rorschach:

Raven's test? 5

I think so, I forgot already... I picked a 5.

The site hangs, I reloaded the test pages several times.

https://xn--80adjkcael4abtflqeskx.xn--p1ai/#start.

found a course there from some uni on deep MO, signed up

[mytarmailS]

Valeriy Yastremskiy:
mytarmailS

thanks

In continuation of this

I used Valery's script to manually select examples for training, which I liked.

Trained Kohonen, pulled interesting prototypes after clustering...

I think it turned out pretty good...

Now I can use these prototypes to look for similarities in the graph...

[[Deleted]]

Rorschach:
ods.ai is interesting, but I don't know how to get there

Send me an email and I'll send you an invitation.

[Aleksey Vyazmikin]

Maxim Dmitrievsky:

What?

If it gives a prediction for a bar or several, it is not enough to make a decision, it is not much better than the MA. I am interested in the deviations from this notional MA for profits.

[Aleksey Vyazmikin]

Maxim Dmitrievsky:

I think so, I forgot already. I chose 5.

The site hangs, I reloaded the test pages several times.

https://xn--80adjkcael4abtflqeskx.xn--p1ai/#start

found a course there from some uni on deep MoD, signed up.

They always discriminate, my region does not get in :(

[Aleksey Vyazmikin]

mytarmailS:

thanks

In continuation of this

Using Valery's script, I manually selected examples for training that I liked.

Trained kohonen, pulled prototypes I found interesting after clustering...

I think it turned out pretty good...

Now I can use these prototypes to look for similarities in the graph...

The clustering doesn't seem very good, lots of self-similar pictures.

And the approach itself is interesting, how many points were used to describe the structure?

[mytarmailS]

Aleksey Vyazmikin:

Clustering seems not very good, a lot of self-similar pictures.

And the approach itself is interesting, how many points were used to describe the structure?

Clustering is very good...

I just want to describe on the level of images what I can't describe by code myself...

It's simple...

the work of an automatic recognizer...

of course it doesn't work )) but I got what I wanted ...