Timeline for What is the reasons for a model to have a high cross validation score and yet underperforms on unseen data?
Current License: CC BY-SA 4.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 8, 2019 at 2:58 | vote | accept | CommunityBot | ||
| Jan 3, 2019 at 18:48 | answer | added | cbeleites | timeline score: 2 | |
| Jan 3, 2019 at 17:56 | comment | added | user217442 | the training set is mix of the data of all users (except one user who serves as a validation set). Thus, the final dataset does not have any label information of which user the data belongs to. | |
| Jan 3, 2019 at 17:49 | comment | added | Sandro | I'm not entirely sure what your input and your output is. One thing that comes to mind: Are you sure that the train-test split is done on the subjects? Otherwise it can happen that your algorithm learns to identify the users and connects it with the labeled information. | |
| Jan 3, 2019 at 17:36 | history | asked | user217442 | CC BY-SA 4.0 |