The adjustment phase guarantees that every cluster has at leastknodes. The clusters
formed in early iterations of the clustering algorithm have lower chances of having
less than knodes compared to the clusters formed towards the end of clustering
phase. This is due to the fact that clusters formed in initial stage of clustering
algorithm have enough nodes to satisfyk-anonymity andl-diversity, but the clusters
formed in the end may not satisfy this die due to scarcity of nodes. The adjustment
phase algorithm starts from the last cluster formed. The clusters having less thank
nodes are merged together so that the resulting cluster has at leastknodes. If there
is only one cluster left with less than k-nodes, its nodes are moved to the cluster
whose nodes have maximum similarity with respect to that node. In the example
given, there is no such cluster which does not satisfyk-anonymity andl-diversity,
so adjustment is not required