Research and Application of Improved CHAMELEON Algorithm Based on Condensed Hierarchical Clustering Method
Dongwei Guo, Jingjing Zhao, Jici Liu · 2019
CHAMELEON clustering algorithm uses dynamic modeling to find high quality clusters of different shapes, sizes and densities. When the CHAMELEON algorithm constructs the k-nearest neighbor graph G, the k value is difficult to select and has a great influence on the result. Using the multi-level partitioned hMetic algorithm for segmenting large hyper graphs to divide Gk is a coarse partition, which easily leads to uncertainty of results. Aiming at this shortcoming, this paper proposes a new condensed hierarchical clustering method based on AGNES algorithm, and uses this convenient and efficient condensed hierarchical clustering method to replace the traditional hMetic algorithm to divide Gk to generate sub-clusters.