Machine learning in trading: theory, models, practice and algo-trading - page 2988

New comment
 
Maxim Dmitrievsky #:
Based on all sorts of tsos and shaplets doesn't make it in the financial markets. Like it or not.

So if you use these methods as complementary to other predictors - maybe they will add a couple of per cent.

 
Aleksey Vyazmikin #:

So if you use these methods as complementary to other predictors - maybe they will add a couple of percentages.

It would be nothing +2% :)
Finding inefficiency is harder than finding a signal or pattern, so it's kind of a no-go here
 
Maxim Dmitrievsky #:
It will be nothing +2% :)

By the way, it might even be worth applying these methods not on price but on predictors, as they can be more stationary.

 
Aleksey Vyazmikin #:

By the way, it may even be worthwhile to apply these methods on predictors rather than price, as they may be more stationary.

Please, but looking for inefficiencies is not looking for signals or patterns. It's something metaphysical. You need another hyperboloid of some sort for that.
 
Maxim Dmitrievsky #:
Please, but looking for inefficiencies is not looking for signals or patterns. It's something metaphysical. You need a different kind of hyperboloid for that.

I guess it all depends on goals and objectives. I don't set goals that require huge budgets. And random discoveries only happen in experiments.

 
Aleksey Vyazmikin #:

So if you use these methods as complementary to other predictors - maybe they'll add a couple of per cent.

For example, if you have a 5% MOE advantage, then add 2% of 5%, for a total of 5.1%. Significantly increasing the calculation time and complexity of the system.

 
In the video he said to feed more different predictors and the MO will sort itself out. Apparently, he has not heard about input rubbish = output rubbish. Recently even Gref quoted this phrase.
But perhaps there is no debris in gas swirls caused by welding defects. So in this particular example all transformations will work for him.
 
Forester #:
In the video he said, , that you need to feed more different predictors and the MO will sort itself out. Apparently he has not heardabout rubbish at the input = rubbish at the output. Recently even Gref quoted this phrase.
But perhaps there is no debris in gas swirls caused by welding defects. So in this particular example all transformations will work for him.

The saddest thing is that you do not even realise what nonsense you are talking....

 
Aleksey Nikolayev #:
Another seasonal aggravation of radio amateurs. Again they catch and decipher signals from the market.

And what digital educational environment is not to your table))))

 
Aleksey Nikolayev #:
Another seasonal aggravation of radio amateurs. Again they catch and decode signals from the market.
Seasonal is nothing. ML amateurs have a year-round aggravation.)
1...298129822983298429852986298729882989299029912992299329942995...3523
New comment
Reason: