226. A-Ad-M
Damped Holt-Winters, multiplicative seasonality
Example:
Given
- Smoothing parameters: , , ,
- Initial states: , , , seasonal period
- 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
Observation:
Conditional mean (one-step-ahead forecast ):
Innovation:
State updates:
Forecast steps ahead from time (using current-period states):
where is the forecast horizon (how many steps ahead); picks the right seasonal slot for the period steps ahead (cycles through ).
Step 2 — apply at
Step 3 — iterate
Each column header is the equation that produced its values. Values rounded to 4 decimal places; arithmetic performed at full precision.
| 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 |