Timeline for auto.arima cannot offer best model lowest aic
Current License: CC BY-SA 3.0
12 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Feb 25, 2017 at 17:26 | vote | accept | Kefei | ||
| Feb 23, 2017 at 23:27 | comment | added | IrishStat | Your SMA coefficient (-.7706 or -.8901) suggests possible unwarranted seasonal differences as it is in the neighborhood of -1.0. | |
| Feb 22, 2017 at 21:18 | answer | added | Rob Hyndman | timeline score: 4 | |
| Feb 22, 2017 at 17:45 | comment | added | Richard Hardy |
If you check the algorithm behind auto.arima, you will see that AIC is used within a fixed order of integration, but not across them. The order of integration (seasonal and simple) is determined by tests (OCSB and KPSS, respectively).
|
|
| Feb 22, 2017 at 17:29 | comment | added | Kefei | i select minimize the aic when using the auto.airma, it should give the model with the lowest aic. :> | |
| S Feb 22, 2017 at 15:58 | history | suggested | Ferdi | CC BY-SA 3.0 |
Include a Hyperlink
|
| Feb 22, 2017 at 15:35 | review | Suggested edits | |||
| S Feb 22, 2017 at 15:58 | |||||
| Feb 22, 2017 at 15:10 | comment | added | Richard Hardy | You cannot directly compare the AIC for models with different integration orders. Thus your AIC comparison does not tell which model is actually better. | |
| Feb 22, 2017 at 15:07 | history | edited | Stefan | CC BY-SA 3.0 |
added 6 characters in body
|
| Feb 22, 2017 at 14:55 | history | edited | Stefan | CC BY-SA 3.0 |
improved formatting
|
| Feb 22, 2017 at 14:51 | review | First posts | |||
| Feb 22, 2017 at 14:53 | |||||
| Feb 22, 2017 at 14:50 | history | asked | Kefei | CC BY-SA 3.0 |