a refinement to the k-anonymity approach which assures that groups of records specified by the same identifiers have sufficient diversity to prevent inferential disclosure
Source(s):
NISTIR 8053
from
Machanavajjhala, J., et al