Timeline for How to fit a regression model for data like this
Current License: CC BY-SA 3.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Nov 6, 2015 at 19:36 | comment | added | Nick Cox | What's the science (e.g. economics) here? Price presumably must be positive; what else do we know? What is generation? Do you expect a turning point (or two)? There is so much scatter here that a little theory even on what is plausible or implausible could help. | |
| Nov 6, 2015 at 16:05 | comment | added | Glen_b |
As I thought; bs by default fits cubic splines, not linear ones. If you do want linear splines, you need to set degree=1 in its arguments
|
|
| Nov 6, 2015 at 16:05 | answer | added | Nether_Nova | timeline score: 1 | |
| Nov 6, 2015 at 15:59 | comment | added | user78229 | Much more informative than a scatterplot of price*generation would be to look at how these variables vary over time. That might suggest leveraging some of the classic approaches to time series, e.g., Holt-Winters decompositions, etc. | |
| Nov 6, 2015 at 15:37 | history | edited | Penguin_Knight | CC BY-SA 3.0 |
edited title
|
| Nov 6, 2015 at 15:22 | comment | added | Nyxynyx | @Glen_b Oops left out the code. Updated the question with it. | |
| Nov 6, 2015 at 15:22 | history | edited | Nyxynyx | CC BY-SA 3.0 |
added 320 characters in body
|
| Nov 6, 2015 at 7:22 | comment | added | Glen_b | what was your code to fit the spline? It doesn't look like a linear spline. | |
| Nov 6, 2015 at 6:12 | history | asked | Nyxynyx | CC BY-SA 3.0 |