How could I get outlier instance with k-means clustering in WEKA?

Question

I have used SimpleKmeans class in WEKA, so I do clustering instance as well. But I have a problem during getting outlier instances.

I supposed, each cluster in this class has a center(or centroid) and a radius, so I could find outliers by checking all clusters' circle with its centroid and its radius. Although I couldn't find any variables or functions that get cluster's radius.

Now, Do you know any other way for finding outliers at SimpleKmeans class in WEKA? Or Any variables that shows radius for each cluster?

Answer 1

I couldn't find any radius variable in SimpleKmeans class, however I use alternative solution.

Clustering find best cluster that has minimum Euclidean distance(or manhattan ,..), so It will outlier instance if the nearest cluster is more than a specific threshold.