How useful are Featrues(but Time & Price itself) in LSTM deep Learning for Trading ? about adv_fin_ml_exercises HOT 6 OPEN

Frac diff can be implemented as a fir filter, lstm can learn the parameters, the point is, instead of feature scaling/manipulation with fracdiff, you are doing a model fitting with lstm (classifier/regressor), and as a model tou must check what it fits, noise or signal, that’s why you use filter and model in two steps, to know exactly ehat each one do

from adv_fin_ml_exercises.

Isn't this the reason why MLDP recommends fractional differentiation?

Having this said I'm not sure about this. Might also depend on the indicators you are looking a. I could imagine that there is value in adding such features but this is a guess. Maybe you simply give it a try and let us know what you find. :-)

from adv_fin_ml_exercises.

Isn't this the reason why MLDP recommends fractional differentiation?

Having this said I'm not sure about this. Might also depend on the indicators you are looking a. I could imagine that there is value in adding such features but this is a guess. Maybe you simply give it a try and let us know what you find. :-)

Recently heard about fractional differentiation, but sadly couldnt find anything useful to understand the concept, like a formula or so on google, could you explain or provide a link to help me understand?

from adv_fin_ml_exercises.

Lstm can reproduce a diff frac, the problem is if it will fit at noise or signal

from adv_fin_ml_exercises.

Maybe this helps:

https://github.com/philipperemy/fractional-differentiation-time-series

from adv_fin_ml_exercises.

Frac diff can be implemented as a fir filter, lstm can learn the parameters, the point is, instead of feature scaling/manipulation with fracdiff, you are doing a model fitting with lstm (classifier/regressor), and as a model tou must check what it fits, noise or signal, that’s why you use filter and model in two steps, to know exactly ehat each one do

So you mean i put this filter over every batch or how do you define filter (in python terms)?

from adv_fin_ml_exercises.