Cumulative match score I have seen loads of graphs in papers of cumulative match scoring, but I can't find any information about what it means, or how it is created.
A context that would be useful to see the explanation with respect to would be an N class (N>>2) KNN pattern classification problem.
EDIT: An example:

From [this][2] paper by Wagg:


*

*Wagg, D.K.; Nixon, M.S., "On automated model-based extraction and analysis of gait," Automatic Face and Gesture Recognition, 2004. Proceedings. Sixth IEEE International Conference on , vol., no., pp.11,16, 17-19 May 2004
doi: 10.1109/AFGR.2004.1301502


It is concerned with gait recognition from a number of subjects. The top of the last page shows an example of the kind of graph I don't understand.
 A: I have finally found the answer.
Considering the classification problem with KNN with M classes and N inputs to predict. The output would be an M by N array of distances from N to M. 
Ordering and decoding the indexes of this gives the actual classification: an M by N matrix of class labels where the first column is the closest label and the Mth column is the furthest label.
Taking the first column of this matrix gives a 1 by N matrix of the "headline" result.
For each row, the cumulative match score is whether there is a correct result within the first R columns. R is called the rank.
The cumulative match characteristic is the sum of these rows.
The graph therefore shows the headline result as the first rank and the graph can never decrease as R increases.
Assuming the KNN class label matrix is nearestN, the correct result is targets_te and N is the number of samples, some Matlab code to plot this:
cumulativeMatch = zeros(N,1);
correctMatrix = zeros(N,1);

%compare nearestN to targets_te
for i=1:N
   correctMatrix = or(correctMatrix,( (nearestN(:,i) == targets_te)));
   cumulativeMatch(i,1) = sum(correctMatrix==1);
end

cumulativeMatch = 100.*(cumulativeMatch ./ N)

plot(cumulativeMatch(:,1));
axis([1,N,0,100]);

