Timeline for Approximate known non-linear function using linear regression
Current License: CC BY-SA 4.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 29, 2019 at 13:45 | comment | added | BLaursen | No specific function or class, except continuous differentiable functions. I was hoping one could do something inspired by a Taylor approximation that uses information about the derivatives of $f$, but is linear in parameters so that OLS can be used. | |
| Dec 27, 2019 at 2:16 | history | edited | kjetil b halvorsen♦ |
edited tags
|
|
| Dec 26, 2019 at 14:57 | comment | added | whuber♦ | Do you have a specific $f$ or class of $f$ in mind? Note that transforming the $y$ variable produces a different model because the random variables $\varepsilon_i$ are no longer added to $f.$ One routine solution to to use splines, not polynomials. | |
| Dec 26, 2019 at 11:04 | history | asked | BLaursen | CC BY-SA 4.0 |