Mercurial > hg > camir-aes2014
annotate toolboxes/FullBNT-1.0.7/netlab3.3/confmat.m @ 0:e9a9cd732c1e tip
first hg version after svn
| author | wolffd |
|---|---|
| date | Tue, 10 Feb 2015 15:05:51 +0000 |
| parents | |
| children |
| rev | line source |
|---|---|
| wolffd@0 | 1 function [C,rate]=confmat(Y,T) |
| wolffd@0 | 2 %CONFMAT Compute a confusion matrix. |
| wolffd@0 | 3 % |
| wolffd@0 | 4 % Description |
| wolffd@0 | 5 % [C, RATE] = CONFMAT(Y, T) computes the confusion matrix C and |
| wolffd@0 | 6 % classification performance RATE for the predictions mat{y} compared |
| wolffd@0 | 7 % with the targets T. The data is assumed to be in a 1-of-N encoding, |
| wolffd@0 | 8 % unless there is just one column, when it is assumed to be a 2 class |
| wolffd@0 | 9 % problem with a 0-1 encoding. Each row of Y and T corresponds to a |
| wolffd@0 | 10 % single example. |
| wolffd@0 | 11 % |
| wolffd@0 | 12 % In the confusion matrix, the rows represent the true classes and the |
| wolffd@0 | 13 % columns the predicted classes. The vector RATE has two entries: the |
| wolffd@0 | 14 % percentage of correct classifications and the total number of correct |
| wolffd@0 | 15 % classifications. |
| wolffd@0 | 16 % |
| wolffd@0 | 17 % See also |
| wolffd@0 | 18 % CONFFIG, DEMTRAIN |
| wolffd@0 | 19 % |
| wolffd@0 | 20 |
| wolffd@0 | 21 % Copyright (c) Ian T Nabney (1996-2001) |
| wolffd@0 | 22 |
| wolffd@0 | 23 [n c]=size(Y); |
| wolffd@0 | 24 [n2 c2]=size(T); |
| wolffd@0 | 25 |
| wolffd@0 | 26 if n~=n2 | c~=c2 |
| wolffd@0 | 27 error('Outputs and targets are different sizes') |
| wolffd@0 | 28 end |
| wolffd@0 | 29 |
| wolffd@0 | 30 if c > 1 |
| wolffd@0 | 31 % Find the winning class assuming 1-of-N encoding |
| wolffd@0 | 32 [maximum Yclass] = max(Y', [], 1); |
| wolffd@0 | 33 |
| wolffd@0 | 34 TL=[1:c]*T'; |
| wolffd@0 | 35 else |
| wolffd@0 | 36 % Assume two classes with 0-1 encoding |
| wolffd@0 | 37 c = 2; |
| wolffd@0 | 38 class2 = find(T > 0.5); |
| wolffd@0 | 39 TL = ones(n, 1); |
| wolffd@0 | 40 TL(class2) = 2; |
| wolffd@0 | 41 class2 = find(Y > 0.5); |
| wolffd@0 | 42 Yclass = ones(n, 1); |
| wolffd@0 | 43 Yclass(class2) = 2; |
| wolffd@0 | 44 end |
| wolffd@0 | 45 |
| wolffd@0 | 46 % Compute |
| wolffd@0 | 47 correct = (Yclass==TL); |
| wolffd@0 | 48 total=sum(sum(correct)); |
| wolffd@0 | 49 rate=[total*100/n total]; |
| wolffd@0 | 50 |
| wolffd@0 | 51 C=zeros(c,c); |
| wolffd@0 | 52 for i=1:c |
| wolffd@0 | 53 for j=1:c |
| wolffd@0 | 54 C(i,j) = sum((Yclass==j).*(TL==i)); |
| wolffd@0 | 55 end |
| wolffd@0 | 56 end |