Timeline for Which ML algorithm can learn non-linear interaction effects?
Current License: CC BY-SA 4.0
2 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Oct 11, 2020 at 8:41 | comment | added | HansHupe | Adding manually different orders of interaction terms require domain knowledge, which I want to minimize, also there is a risk of overfitting when interaction terms are added between all variables. I am kind of disappointed that ML algorithms are not able to learn that effects. At the moment I am experimenting with MARS. | |
| Oct 11, 2020 at 5:27 | history | answered | hehe | CC BY-SA 4.0 |