How can I evaluate performance of a density based clustering algorithm? Are there any adopted metrics of evaluation?
In a recent study on density based clustering method (Yading) they have used the following measure for evaluation.
Normalized Mutual Information (NMI): Danon L, D´ıaz-Guilera A, Duch J and Arenas A 2005 J. Stat. Mech. P09008
OR
Refer to the following link
http://nlp.stanford.edu/IR-book/html/htmledition/evaluation-of-clustering-1.html
http://www.vldb.org/pvldb/vol8/p473-ding.pdf
Whether it is density based, partitioning based etc, you can use internal or external evaluation approaches to check the performance of your clustering algorithm. You can read more here:
https://en.wikipedia.org/wiki/Cluster_analysis
Also, you can read this article for cluster path lengths approach.
https://hal.archives-ouvertes.fr/inria-00542690/PDF/icdm2010ClusterPathCamera.pdf
All the best.
Different indexing measures are there such as DBINDEX, DUNNINDEX, CSINDEX.
...For this purpose is possible to use several methods: Neural Networks, Multivariate Methods, Factorial Methods, Statistical diagnostic test, etc. But firstly is necessary obtains very clear information as input of all method mentioned in order to get logic results.
- Badges
- Science topic
- Similar topics
- Computer Science
- Data Mining