252. ARMA
Autoregressive moving average
Combines AR and MA structure. Requires stationary input series.
Parameters: , , ,
Orders: ,
Example:
Given
- Orders: ,
- Parameters: , ,
- Initial conditions: ,
- Data:
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | |
| 12 | 10 | 8 | 11 | 14 | 12 | 9 | 13 | 16 | 14 | 11 | 15 | 18 | 16 | 13 | 17 |
Step 1 — formula
Substitute , into the ARMA recursion:
Forecast (set ):
Innovation:
Step 2 — apply at
Plug in , , , , :
Step 3 — iterate
Each row uses the previous observation and the previous innovation . Values rounded to 4 decimal places.
| 1 | 12 | ||
| 2 | 10 | ||
| 3 | 8 | ||
| 4 | 11 | ||
| 5 | 14 | ||
| 6 | 12 | ||
| 7 | 9 | ||
| 8 | 13 | ||
| 9 | 16 | ||
| 10 | 14 | ||
| 11 | 11 | ||
| 12 | 15 | ||
| 13 | 18 | ||
| 14 | 16 | ||
| 15 | 13 | ||
| 16 | 17 | ||
| 17 | — | — |