Timeline for Forward or backward sequential feature selection?
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| May 27, 2019 at 6:23 | comment | added | beginner_ | @gung Agree about backward/forward selection not being needed. But OP could try (or could have tried) low variance filtering and correlation filtering of features. | |
| Jan 5, 2018 at 22:33 | vote | accept | prashanth | ||
| Jan 5, 2018 at 22:26 | comment | added | gung - Reinstate Monica | @prashanth, I would need to know a lot more about your project, but their recommendation sounds mindless. (Not that I'm skeptical--I see that plenty.) For your case (p=22, k=5000), there is little need for feature selection for anything. If you want to appease the reviewers, something like regularization or PCA (both listed on the MATLAB documentation page they linked to) would be more defensible. | |
| Jan 5, 2018 at 22:21 | comment | added | prashanth | Thank you very much. I wanted to try the feature selection method based on the reviewer comments for a research article. Please see the modified question to see the comment. The reviewer had suggested to do feature selection. I guess now i can use the above references for response. | |
| Jan 5, 2018 at 21:57 | history | answered | gung - Reinstate Monica | CC BY-SA 3.0 |