Counting How Many Cluster Index

I am doing clustering with k-means, and here is the code:

[g c] = kmeans(data,k,'dist','SqEuclidean');

g is the cluster index

c is the centroid

k is the number of clusters, I input '3' in it.

When the code is executed, here is the sample result:

40    60    69     1
50    75    78     2
65    70    75     2
79    80    72     3
88    80    80     3
36    67    66     1
40    60    69     1
77    85    65     3
76    77    79     3

My question is, how to count the cluster index for each cluster? What script I should write to show this thing:

"Cluster1 has three members

Cluster2 has two members

Cluster3 has four members"

I'm getting confused because the cluster index only have one variable. Any idea?

Thank you.

Answer 1

[Tom Lane]

Look at these, and see if it helps:

unique(g)
find(g==3)

[Alvi Syahrin]

Thank you. Anyway, it is working with accumarray function. ^_^