Why do I get this message : Error using kmeans ---X must have more rows than the number of clusters.

Hello,

When I run a collegue script, i get this error message: Error using kmeans ---X must have more rows than the number of clusters. What are my option to solve this problem?

Thank you.

Andree

Answer 1

[Kelly Kearney]

How many data points do you have? And how many clusters are you attempting to create? You can't group 4 points into 5 clusters, for example.

Answer 2

[Andree-Sylvie Carbonneau]

This is the code:

ab = IM1; 
nrows = size(ab,1);
ncols = size(ab,2);
ab = reshape(ab,nrows*ncols,1);

nColors = 3; 
[cluster_idx cluster_center] = kmeans(ab,nColors,'distance','sqEuclidean',...
    'Replicates',nColors + 1,'display','iter');
clear ab nColors

[José-Luis]

size(ab)?

[Andree-Sylvie Carbonneau]

Size(ab,1)... is from the Color-Based Segmentation Using K-Means Clustering example that I found on MATHWORKS documentation center.

Answer 3

[Thulfiqar Hussein]

the value (X/K) should be integer number without fractions. for example if you have 10 objects and cluster them to 3 groups you will get 10/3 = 3.3 and this error will pop up, but if you cluster them to 2 or 5 groups there will be no problem.

Answer 4

[Laura Hensel]

You should also check if you have any missing values in your matrix