I am clustering based on my cosine similarity matrix. Now I want to plot/visualize my very large clusters. I am interested in using a tool that is better than sklearn.
Please recommend me.
$\begingroup$
$\endgroup$
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
1
ELKI has some very nice cluster visualizations. You could also use tSNE or MDS.
But as you used cosine, I assume you have text data, and that is not easy to visualize. You probably first need to figure out how to visualize your input data, then you can think about adding cluster information to this.
answered Sep 28, 2017 at 20:10
-
$\begingroup$ Thanks a lot for the great answer. I actually have a feature vectors
x = [0.7, 0.2, 0, 0.1]like this. I use these feature vectors to calculate the cosine distance. Can I use these feature vectors directly to plot the graphs in ELKI tool? $\endgroup$– VolkaSep 29, 2017 at 0:07