Table 2 |
||||||||
|
Mean absolute relative error of out of series prediction at forecasting horizons of 1 to 4 months ahead for districts in Sri Lanka for Holt Winters models. |
||||||||
| District |
Horizon 1 |
Horizon 2 |
Horizon 3 |
Horizon 4 |
||||
|
|
||||||||
| Model |
H |
HW |
H |
HW |
H |
HW |
H |
HW |
|
|
||||||||
| Ampara |
0.43 |
0.39 |
0.65 |
0.52 |
0.83 |
0.63 |
0.86 |
0.67 |
| Anuradhapura |
0.34 |
0.22 |
0.66 |
0.35 |
0.99 |
0.45 |
1.22 |
0.53 |
| Badulla |
0.46 |
0.54 |
0.67 |
0.75 |
0.87 |
0.95 |
0.84 |
0.96 |
| Batticaloa |
0.41 |
0.41 |
0.65 |
0.65 |
0.82 |
0.82 |
0.97 |
0.97 |
| Colombo |
0.35 |
0.37 |
0.39 |
0.43 |
0.44 |
0.48 |
0.46 |
0.53 |
| Galle |
0.50 |
0.61 |
0.59 |
0.74 |
0.67 |
0.83 |
0.79 |
0.96 |
| Gampaha |
0.43 |
0.43 |
0.59 |
0.59 |
0.70 |
0.70 |
0.78 |
0.78 |
| Hambantota |
0.36 |
0.36 |
0.57 |
0.56 |
0.76 |
0.73 |
0.88 |
0.87 |
| Jaffna |
0.43 |
0.46 |
0.62 |
0.63 |
0.79 |
0.85 |
0.85 |
0.97 |
| Kalutara |
0.55 |
0.61 |
0.72 |
0.80 |
0.81 |
0.91 |
0.88 |
0.97 |
| Kandy |
0.37 |
0.37 |
0.50 |
0.50 |
0.56 |
0.57 |
0.57 |
0.57 |
| Kegalle |
0.39 |
0.40 |
0.63 |
0.62 |
0.83 |
0.82 |
0.94 |
0.95 |
| Kilinochchi |
0.58 |
0.60 |
1.08 |
1.12 |
2.50 |
2.26 |
2.70 |
2.17 |
| Kurunegala |
0.34 |
0.26 |
0.61 |
0.43 |
0.76 |
0.57 |
0.85 |
0.70 |
| Mannar |
1.41 |
1.57 |
1.74 |
1.98 |
1.61 |
2.63 |
1.78 |
2.28 |
| Matale |
0.45 |
0.41 |
0.73 |
0.63 |
0.96 |
0.74 |
1.13 |
0.81 |
| Matara |
0.37 |
0.35 |
0.42 |
0.40 |
0.49 |
0.48 |
0.52 |
0.52 |
| Moneragala |
0.31 |
0.31 |
0.42 |
0.41 |
0.54 |
0.52 |
0.62 |
0.63 |
| Mullaitivu |
1.08 |
1.19 |
1.73 |
1.70 |
2.21 |
2.54 |
2.73 |
2.38 |
| Nuwara Eliya |
0.49 |
0.50 |
0.61 |
0.60 |
0.69 |
0.69 |
0.69 |
0.69 |
| Polonnaruwa |
0.37 |
0.37 |
0.60 |
0.60 |
0.76 |
0.76 |
0.82 |
0.82 |
| Puttalam |
0.42 |
0.37 |
0.67 |
0.49 |
0.88 |
0.64 |
1.00 |
0.76 |
| Ratnapura |
0.36 |
0.31 |
0.55 |
0.47 |
0.64 |
0.56 |
0.74 |
0.66 |
| Trincomalee |
0.53 |
0.56 |
0.82 |
0.75 |
1.15 |
0.97 |
1.35 |
1.07 |
| Vavuniya |
1.89 |
2.02 |
2.82 |
3.93 |
2.45 |
14.21 |
2.19 |
4.11 |
|
|
||||||||
|
H = Holt's two parameter exponential smoothing; HW = Holt-Winters three parameter exponential smoothing (including seasonality). Values in bold italic represent a better mare as compared to the best (S)ARIMA model (without rainfall). |
||||||||
|
Briët et al. Malaria Journal 2008 7:76 doi:10.1186/1475-2875-7-76 |
||||||||