Timeline for Why is my high degree polynomial regression model suddenly unfit for the data?
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jun 8, 2017 at 9:35 | comment | added | kjetil b halvorsen♦ | High-order polynomial very seldom are good models, polynomials are way to stiff. Try splines! search this site for splines. | |
| Jul 28, 2016 at 15:51 | comment | added | Near | @KevinZhang What is the optimisation algorithm that you're using? | |
| Jul 28, 2016 at 15:49 | comment | added | Kevin Zhang | Sorry I can't, it's proprietary data | |
| Jul 28, 2016 at 15:47 | comment | added | Near | @KevinZhang Can you give me a link to the dataset? | |
| Jul 28, 2016 at 14:10 | comment | added | Kevin Zhang | My train and test set errors are both huge when I increase the order polynomial, so this doesn't seem to be about overfitting. | |
| Jul 28, 2016 at 7:35 | history | edited | Near | CC BY-SA 3.0 |
Elaborated the answer
|
| Jul 28, 2016 at 3:45 | comment | added | Matt Krause | Hi! Welcome to Cross Validated! Could I push you to elaborate a little bit more in your answer? I think you might be on to something, though it does seem odd for the training error to drop if the issue is overfitting. | |
| Jul 28, 2016 at 3:08 | review | First posts | |||
| Jul 28, 2016 at 3:45 | |||||
| Jul 28, 2016 at 3:03 | history | answered | Near | CC BY-SA 3.0 |