How to fix appropriate radius(r) for the significance test in Kth nearest neighbor clustering?

[partha1]

What value of radius(r) is to be choosen for the significance test in Kth nearest neighbor clustering ? (Clustering variables are measured in same scale and hence, no standardization is required in my data).

[SteveDenham]

This is pretty much a decision based on knowledge about the nature of the data.  You have to make a choice as to an appropriate value, combining your need for successful discrimination with the anticipated granularity of the clusters.  Arbitrary values make little sense.

Steve Denham