Chameleon is a hierarchical clustering algorithm that overcomes the limitations of the existing models and the methods present in the data warehousing. This method operates on the sparse graph having nodes: that represent the data items, and edges: representing the weights of the data items.

This representation allows large dataset to be created and operated successfully. The method finds the clusters that are used in the dataset using two phase algorithm.

☛ The first phase consists of the graph partitioning that allows the clustering of the data items into large number of sub-clusters.

☛ Second phase uses an agglomerative hierarchical clustering algorithm to search for the clusters that are genuine and can be combined together with the sub-clusters that are produced.