Timeline for Why exactly using a test set for model evaluation is a bad idea?
Current License: CC BY-SA 3.0
6 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Sep 28, 2017 at 8:25 | vote | accept | Bitcoin Cash - ADA enthusiast | ||
| Sep 26, 2017 at 7:14 | answer | added | Neil Slater | timeline score: 2 | |
| Sep 25, 2017 at 23:38 | comment | added | Alex | Usually you train the model and every N iterations (e.g. 1000) test it on a separate validation set. Validation set error is often used to stop the training. Test set is used only once after you are done training, e.g. to compare the efficiency of different algorithms. | |
| Sep 25, 2017 at 22:16 | answer | added | noe | timeline score: 2 | |
| Sep 25, 2017 at 21:59 | history | edited | Bitcoin Cash - ADA enthusiast | CC BY-SA 3.0 |
added italic
|
| Sep 25, 2017 at 21:43 | history | asked | Bitcoin Cash - ADA enthusiast | CC BY-SA 3.0 |