a question about CLUTO clustering vs. kmeans
Hi, all,
I'm trying to cluster the following dataset into two clusters:
[row col] = [10 2]
0.5 0.6
0.6 0.7
0.9 0.9
4.0 4.0
4.5 4.5
4.0 4.1
3.6 3.8
4.2 5.0
4.1 4.6
1.0 1.0
And CLUTO gives the result as:
1
1
0
0
0
0
0
1
1
0
If we plot the data, we can easily notice that the result should be as follows, which is the same result using kmeans method.
1
1
1
0
0
0
0
0
0
1
So how can I use CLUTO to achieve the same result with kmeans'?
Looking forward to your replies. Thanks a lot.
Submitted by lxduan on Sun, 2008-09-07 05:19
»
- Login to post comments
RE: Cluto's default similarity
Cluto's default similarity function is based on cosine and this is why the results do not look as you expected them to look. Try using the graph-based clustering with euclidean distance.