Agglomerative Hierarchical Clustering Based on Local Optimization for Cluster Validity Measures

摘要

Modularity is an evaluation measure for graph clustering. Louvain method is constructed by local optimization for modularity and is bottom up method as well as agglomerative hierarchical clustering. Cluster validity measures are used to evaluate cluster partitions as well as modularity. They are traditional evaluation measures in the field of clustering. We propose a novel graph clustering which is based on agglomerative hierarchical clustering. The proposed method in this study is constructed by local optimization for cluster validity measures. The effectiveness of the proposed method is shown through numerical examples. Numerical examples show that the proposed method has different clustering propety from Louvain method because of the feature of cluster validity measures.