h
|
1
|
2
|
3
|
n
h
|
A
h
|
W
h
|
Ï•
|
---|
1
| 51 | 23 | 13 | 87 | 22,044 | 0.0032 | 0.0033 |
2
| 0 | 416 | 15 | 431 | 2,694,787 | 0.397 | 0.41 |
3
| 1 | 20 | 410 | 431 | 4,071,576 | 0.600 | 0.58 |
Total | 52 | 459 | 438 | 949 | 6,788,387 | 1 | 1 |
h
|
A
h, a
| S(A
h, a
) |
Lower CI
|
Upper CI
|
User's
|
Prod's
| Prop. |
1
| 22,370 | 9,519 | 3,331 | 41,408 | 59 | 98 | 0.19 |
2
| 2,795,765 | 47,690 | 2,700,384 | 2,891,146 | 97 | 91 | 38 |
3
| 3,970,273 | 48,528 | 3,873,217 | 4,067,328 | 95 | 94 | 57 |
 |
Overall accuracy
|
Area-weighted accuracy
| Â | Â | Â | Â | Â |
 | 92 | 96 |  |  |  |  |  |
- *Map categories are rows (1 = forest-to-non-forest; 2 = stable forest; 3 = stable non-forest) while the true categories are in columns. h is the map category; n
h
is the number of samples in h; A
h
is the area of map category h, with A
h, a
being the adjusted value of A
h
; W
h
is the proportional area of h; Ï• is proportion of samples in map category h; S(A
h, a
) is the standard error of the adjusted area; Lower/Upper CI are the upper and lower confidence intervals; User's and Prod's are user's and producer's accuracy in percent; Prop. is the user's accuracy weighted by map category area proportion. Overall accuracy and Area-weighted accuracy are expressed in percent. Area and confidence intervals are expressed in hectares.