Chameleon is a hierarchical clustering technique that overcomes the constraints of current Data Warehousing concepts and methodologies. This approach works with a sparse graph with nodes that represent data items and edges that indicate data item weights.
This format enables the creation and operation of huge datasets. Using a two-phase approach, the method discovers the clusters that are utilized in the dataset.
The first phase is graph partitioning, which allows data items to be clustered into a large number of sub-clusters.
The second step employs an agglomerative hierarchical clustering method to look for real clusters that may be merged with the sub-clusters that are generated.