Machine learning in trading: theory, models, practice and algo-trading - page 2880

New comment
 
Happy New Year! New wishes and fulfilment of the deepest desires!
 
Vladimir Perervenko #:

Traits (predictors, variables, functions, features)are individual measurable properties or characteristics of an observed phenomenon. For example, quotes, day of the week, time of day, etc. In our case, as a rule, these are time series .

There is no need to redefine established concepts unnecessarily. This leads to confusion and misunderstanding of the purpose of the discussion. In your case: We are talking about an arbitrary length of a vector of consecutive values of a predictor(trait, function, feature, etc.).Hence the question about columns.

More carefully.

I understand that we need to distinguish more carefully between the length of columns and the number of columns. The former is equivalent to the number of rows, the latter to the length of rows).

Although, with different length of rows the concept of columns is somewhat blurred - instead of a rectangular table (dataframe) the concept of a list of rows becomes relevant.

PS. I also suggested for simplicity to limit ourselves to the case when all signs are only prices on the previous section (with different length)

 
For lSTM, one must be careful to distinguish columns from rows, nested in the 3rd dimension of the sequence length, nested in the 4th dimension, which divides into batches
 
Aleksey Nikolayev #:

I understand that we need to distinguish more carefully between the length of columns and the number of columns. The former is equivalent to the number of rows, the latter to the length of rows)

Although, with different row lengths, the concept of columns becomes somewhat blurred - instead of a rectangular table (dataframe), the concept of a list of rows becomes relevant.

PS. I also suggested for simplicity to limit ourselves to the case when all signs are only prices on the previous plot (with different lengths).

Alexey, you'd better describe what data and what result is expected...
It's hard to think abstractly.
 
Maxim Dmitrievsky #:
That's what a total gypsy liar says, which you are. Don't flatter yourself.

You can tell who is a liar by the number of productive posts. All your posts are not backed up by research.

I can only respect people who speak on the matter and back up their words with research.

I have great respect for Aleksey Vyazmikin and others. He bravely brings out ideas for discussion and shows the result.

Now you know who you really are? Maksimka.)))))

 
Uladzimir Izerski #:

Which one of us is a liar is evident from the number of result posts. All your posts are not backed up by research.

I can only respect people who speak on the matter and back up their words with research.

I have great respect for Aleksey Vyazmikin and others. He bravely brings out ideas for discussion and shows the result.

Now you know who you really are? Maximka.)))

You can respect anyone you want, it won't change your status.
And you shouldn't use diminutive, because I'll beat you in the ring and other events.
 
Maxim Dmitrievsky #:
You can respect anyone, it won't change your status.
.
And you shouldn't diminutive, because I'll beat you in the ring and other events.

I see what you're good at.)

But you're weak when it comes to understanding markets. That's what we're talking about.

 
Uladzimir Izerski #:

I see what you're good at:)

But in understanding markets, you're weak. That's what we're talking about.

Well, you do realise you don't have the natural intelligence to make such judgements.
 
Maxim Dmitrievsky #:
Well you do realise that you lack the natural intelligence to make such judgements

After such a judgement, I am silent.)))))

 
Uladzimir Izerski #:

After such a conclusion, I am silent.)))))

You would have shut up a long time ago if you were a little more intelligent
1...287328742875287628772878287928802881288228832884288528862887...3656
New comment