Quality measures

Any kind of clustering or classification needs an objective measure of its quality. Although we implemented in the Toolbox most conventional indices associated with FCM, including Partition Coefficient, Partition Entropy and Proportion Exponent, our tests showed that the index Lratio was superior to classic FCM indices and so was implemented as the default quality measure in our Toolbox. In vast majority cases values below 5 (to be used as a threshold) means good and acceptable quality of cluster. Values grater than 5 usually associated with poor clusters (regarding their compactness and distance from other clusters in the dataset).

Exact Lratio value for each cluster is shown at the graph legend.

See here for more details: doi:10.1016/j.neuroscience.2004.09.066

Site Map | Printable View | © 2010 - 2019 Dr. Andriy Oliynyk

Creative Commons License