analyzing data - accuracy assessment

[cmlad]

I am analyzing data and would like to perform an accuracy assessment -
computing a error (confusion) matrix, and accuracy statistics
(producer's/user's/confidence interval).
Can I do this in excel?

 

[KC]

Can you provide some data sample as what it should look like? that would help
providing what you want.

-kc

 

[cmlad]

Here you go, "Id" is the reference data (truth) and I am comparing it to 3
different classifications

desc_ Id FR_ISA_EST_6 FR_2001_ISA_7 FR_1997RD_BUF_8
2 1 2 2
2 2 2 2
2 2 2 2
surface 1 1 1 1
road 1 1 1 1
2 1 2 2
2 1 2 2
unknown 1 1 1 1
2 2 2 2
road 1 1 2 1
2 1 2 2
2 1 2 2
2 1 1 2
2 1 2 2
2 1 2 2
2 1 2 2
2 1 2 1
2 1 2 2
surface 1 1 2 1
bldg 1 1 1 1
2 1 2 2
2 1 2 2
2 1 2 2
2 1 2 2
bldg 1 1 1 1
2 2 2 2
road 1 1 2 1
2 2 2 2
2 1 2 2
road 1 1 2 1
surface 1 1 2 1
2 1 2 2
2 1 2 2
2 1 2 1
2 2 2 2
surface 1 1 1 2
2 2 2 2
2 1 2 2
2 1 2 1
2 1 2 2
2 1 2 2
2 2 2 2
2 1 2 2
2 1 2 2
bldg 1 1 1 1
road 1 1 1 1
rail 1 1 1 1
road 1 1 2 1

this is what I have hand calculated (from the entire data set):
FR_1997_RD
Classified class 1 bldg rail road surface unknown class2 Total Users
662 class 1 554 226 780 71.03%
133 bldg 115
9 rail 8
379 road 343
136 surface 84
5 unknown 4
1297 class 2 108 18 1 36 52 1 1071 1179 90.84%
Total 662 133 9 379 136 5 1297
Producers 83.69% 82.58%