File Exchange

image thumbnail

Clustering results measurement

version 1.2 (1.61 KB) by

Measure percentage of Accuracy and the Rand index of clustering results

12 Downloads

Updated

View License

Measure percentage of Accuracy and the Rand index of clustering results
The number of class must equal to the number cluster
Output
Acc = Accuracy of clustering results
rand_index = Rand's Index, measure an agreement of the clustering results
match = 2xk matrix which are the best match of the Target and clustering results
Input
T = 1xn target index
idx =1xn matrix of the clustering results

EX:
X=[randn(200,2);randn(200,2)+6,;[randn(200,1)+12,randn(200,1)]]; T=[ones(200,1);ones(200,1).*2;ones(200,1).*3];
idx=kmeans(X,3,'emptyaction','singleton','Replicates',5);
[Acc,rand_index,match]=AccMeasure(T,idx)

Comments and Ratings (4)

@roslan armina, replace Acc_measure with AccMeasure will work. It's my fault for the example code in the description. It's fixed now.

@Pedro, thank you for pointing out. It's fixed now.

when i use your example, this error come out Undefined function or method 'Acc_measure' for input arguments of type 'double'.
how to solve it? thanks

Pedro

Pedro (view profile)

there is a bug in your code since AccMeasure([1 1 1],[2 2 1]) and AccMeasure([2 2 2],[1 1 2]) give different accuracies

Updates

1.2

1. Bug fix.
2. Corrected the example code.

1.1

correct the example code

MATLAB Release
MATLAB 7 (R14)

Download apps, toolboxes, and other File Exchange content using Add-On Explorer in MATLAB.

» Watch video