Timeline for Is Creating Feature Variation Helps The Model?
Current License: CC BY-SA 3.0
3 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 14, 2016 at 17:56 | comment | added | Hadar | Let me give an example: In particle physics I can have several features to describe an event, each feature on its own is insufficient to discriminate signal from background, but a specific combination (==engineered feature) will output what is called transverse mass which is a very powerful feature. So if i'll randomly generate new features from combination of the original set, maybe I can find a new engineered feature? | |
| Nov 14, 2016 at 17:38 | review | First posts | |||
| Nov 14, 2016 at 17:43 | |||||
| Nov 14, 2016 at 17:34 | history | answered | Thibaut Loiseleur | CC BY-SA 3.0 |