I was surprised to see that with the same processing a linear model outperforms lightGBM. Was wondering if this was the case for others and if you have found that this was still the case after feature engineering?
Thanks!
Hi @sparked!
I haven't done any feature engineering yet.
I have just completed a model selection process yesterday
using the raw data with missing values filled with zeroes.
And I also noticed that linear models outperforms trees and
ensemble models. I could say that it is due to the existence of a
linear relationship between close price and open, high and low prices ...
https://imgur.com/a/OxdqYuT
Yeah. Linear model, and KNeighborsRegressor will output any tree and ensemble model.
'%3e%3cpath fill='%23001489' d='M0 0v6h9V0z'/%3e%3cpath fill='%23e03c31' d='M0 0v3h9V0z'/%3e%3cg stroke='%23fff' stroke-width='2'%3e%3cpath id='d' d='m0 0 4.5 3L0 6m4.5-3H9'/%3e%3cuse xlink:href='%23b' stroke='%23ffb81c' clip-path='url(%23c)'/%3e%3c/g%3e%3cuse xlink:href='%23d' fill='none' stroke='%23007749' stroke-width='1.2'/%3e%3c/g%3e%3c/svg%3e){kind=small}
Hi @sparked!
I haven't done any feature engineering yet.
I have just completed a model selection process yesterday
using the raw data with missing values filled with zeroes.
And I also noticed that linear models outperforms trees and
ensemble models. I could say that it is due to the existence of a
linear relationship between close price and open, high and low prices ...
https://imgur.com/a/OxdqYuT
Yeah. Linear model, and KNeighborsRegressor will output any tree and ensemble model.