Machine learning in trading: theory, models, practice and algo-trading - page 2447

New comment
 
elibrarius:

What kind of mushrooms do you eat?

Any pulse in a large set will create wave phenomena. Modulation is applicable.
 
mytarmailS:

It won't work.

if you dig a little deeper, it might work.

 
Rorschach:

if you dig a little deeper, it might work

For example
 

Check out this entry. Pretty realistic points being made!


 
mytarmailS #:
For example

Fourier/Wavelet


My method.

If I "see" better than others, is there anything I can find?

 
Rorschach #:

Fourier/Wavelet


My method.

If I "see" better than others, is there anything I can find?

I don't think so, because for good earnings it is necessary to see the future, or at least its probability with high probability of fulfillment. As for the rest, even if you have a super smooth olgaritm it is of little use :-(.
 
Rorschach #:

Fourier/Wavelet

My method.

If I "see" better than others, can I find something?

No. No matter how you "look" at the spectra, there is no repeatability in the subsequent price movements after the current pattern, spectral or otherwise... So the pattern must either become more complicated, for example by increasing its size and face the "curse of dimensionality" (read wiki) or solve the problem of invariance of TF, that is to be able to transform data so as to get the same pattern in both a minute and a week, the ach ach signal may be useful here by the way...
There are quite a few other ways, I'm going all at once
 
The cluster analysis of volumes is just one of these ways. There the data on M1 and D1 are the same, or rather patterns. And this makes this analysis advantageous....
 

To look for patterns, you must first master at least the most elementary ways of transforming data...

The brain does this without thinking, but the algorithm needs to prescribe it

================

One, a pattern, and how many degrees of freedom, I haven't even removed the trend, rotated, etc.....

Now do you understand how stupid it is to feed the last n values to AMO and wait for a miracle?

Although I do so, I repent, because it is the easiest way, and do not need to think, and the result is predictable...

 
Mihail Marchukajtes #:

Check out this entry. Pretty realistic points being made!


Divorced. "Pie salesman" -- watch from 7:30 p.m. -- that's what he's talking about himself.

Well, the nonsense called "efficient market" -- he doesn't know what it's about at all. He doesn't understand that this fraud (the theory of the "efficient market") is in and of itself an instrument of bribery. He is not able to understand it.

1...244024412442244324442445244624472448244924502451245224532454...3516
New comment
Reason: