It is based on a paper by Jon Kleinberg from 2002. The paper mentions that an ideal clustering algorithm should have three properties.

1. Scale-invariance: Changing the scale of data shouldn't change the clustering algorithm's output

Comments