Timeline for How to perform seasonal adjustment to a time series?
Current License: CC BY-SA 3.0
4 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 3, 2014 at 6:49 | history | edited | javlacalle | CC BY-SA 3.0 |
I have simplied the answer (only one regression as example) since more relevant information has been given in comments.
|
| Jul 2, 2014 at 20:58 | comment | added | javlacalle | @whuber Very interesting. Knowing the context of the data is always helpful to propose a good model. | |
| Jul 2, 2014 at 20:17 | comment | added | whuber♦ | If you contemplate how heating actually works, you would modify your model and find it can be a much better fit. There will be a temperature threshold above which no heating takes place: that needs to be explicitly in the model (which makes its fitting a nonlinear regression exercise). For temperatures near that threshold, some heating will occur due to daily variations in temperature: this causes the apparent quadratic effect, but it's localized near the threshold temperature. When you account for these in your model, I expect you will find little or no need to introduce seasonal dummies. | |
| Jul 2, 2014 at 20:06 | history | answered | javlacalle | CC BY-SA 3.0 |