Machine learning in trading: theory, models, practice and algo-trading - page 2720

[Aleksey Nikolayev]

Maxim Dmitrievsky #:

And the non-time series thing is just more bullshit with no evidence.

Still, there is a common sense in it. Our initial time is continuous, while in time series it is discrete (by definition).

Substantive work is possible only with discrete time. But transition to discrete time is possible in different ways and it is not necessary to do it only in equal time intervals. It is quite possible to do it by any custom "events". Personally I am closer to the term "Markov moment", because it reflects the main thing - absence of looking into the future. But if someone wants to call it an "event", he is welcome to - as long as he doesn't get confused in his explanations.

[Maxim Dmitrievsky]

Aleksey Nikolayev #:

Still, there is a common sense in it. In our case, the initial time is continuous, while in time series it is discrete (by definition).

Substantive work is possible only with discrete time. But transition to discrete time is possible in different ways and it is not necessary to do it only in equal time intervals. It is quite possible to do it by any custom "events". Personally I am closer to the term "Markov moment", because it reflects the main thing - absence of looking into the future. But if someone wants to call it an "event", he is welcome to - as long as he doesn't get confused in his explanations.

Well, there is discretisation by time-frames, why complicate it. I made custom ones by "CONDITIONS" or "EVENTS", where there is a bold font from the phone, the same bullshit

Just like the Prada books and other ways.

I don't know how to make the case that any oversampling of vp does nothing in terms of improving prediction. Maybe I'm wrong.

[Maxim Dmitrievsky]

Aleksey Vyazmikin #:

Eh, so I won't have a competitive advantage again, sadness.....

It's how you apply it, because there are always options.

[Maxim Dmitrievsky]

There's also market time, it's kind of discrete, according to Mandelsprot, equated to an event - a tick

[Aleksey Nikolayev]

Maxim Dmitrievsky #:
Well, there is discretisation by time-frames, why complicate it. I made custom ones by "CONDITIONS" or "EVENTS", where there is a bold font from the phone, the same bullshit

Just like the Prada books and other methods.

I don't know how to make the case that any oversampling of vr does nothing in terms of improving prediction. I may be wrong.

In general, I agree. Any meaningful discretisation is unlikely to improve much without looking into the future. It's just a matter of personal preference.

[Maxim Kuznetsov]

It is possible that all the vacillation is due to the uncertainty of "what should be the result" - everyone understands the goal differently. There is no goal, there are no intermediate results. There's an exchange of opinions and a search for meaning

---

and the original message in the thread, let me quote:

Machine learning in trading: theory, practice, trading and more.

he is about almost everything ;-) I haven't heard about the absolute "MO is applied in this particular way and no other way, everything else is heresy".

[mytarmailS]

If you can't even discredit it by discrediting it, what have you been doing for years, it's elementary

[mytarmailS]

Another Awesome Video

h ttps:// youtu.be/Zk7dvwrVE0k

[Maxim Dmitrievsky]

mytarmailS #:
Another Awesome Video

h ttps:// youtu.be/Zk7dvwrVE0k

Typical IT evangelist, warms up the subscriber base with an ingratiating voice.

Expands his target audience with references to faith in order to create recognition and trust in the guided masses, and then addiction.

If some people think it's classy, it's worth it.

[mytarmailS]

Maxim Dmitrievsky #:
Typical IT evangelist, warms up the subscriber base with an ingratiating voice

Expands his target audience with references to faith in order to create recognition and trust in the slave masses, and then addiction.

If some people think it's classy, it's worth it.

You have no idea what convolutional networks are, let alone the principles of strong AI, what you can appreciate at all, sit quietly, fap on your catbusters and don't embarrass yourself.