• |
|
|
|
|
|
|
| = | S |
| + | 1 | T |
| ||
|
|
|
|
|
|
|
|
| t + 1 | t | t | |||||
|
|
|
| S1S3 | – S22 | |
|
|
|
| α |
| ||||
c = exp |
| 1 | |
|
|
|
|
|
|
|
|
| ||||
+ S3 |
|
|
|
|
|
|
|
|
| |||||||
|
| n | S1 | – 2S2 |
|
|
|
|
|
|
|
|
| |||
• |
|
|
|
|
|
|
|
|
|
|
|
|
|
| ||
a | = exp |
| (b – 1)(S2 – S1) | |
| ||||
2 | ||||
|
|
| b(bn – 1) | |
|
|
|
| |
•Where S1, S2, and S3 are:
n
S1 = ∑Inyi = n lnc +
i = 1
•
2n
bn – 1 b( lna)
b – 1
S |
| = | ∑ | Iny |
| = n lnc + b | n + 1 | ( lna) | bn | – 1 |
2 | i |
| ||||||||
|
|
|
|
|
| b | – 1 | |||
i = n + 1
•
•a, b and c are determined by solving the three equations above simulta- neously.
Forecasting With Exponential Smoothing
•a = smoothing constant (0 < a < 1)
•Xt = actual current period usage
