Document clustering based on vector quantization and growing-cell structure

In this paper, we proposed a new hybrid clustering algorithm based on Vector Quantization (VQ) and Growing-Cell Structure (GCS). The basic idea is using VQ to refine the GCS clustering results and thus to improve the clustering performance. Moreover, the output of the proposed clustering algorithm has a graph structure which is generated gradually during the incremental self-learning process. We evaluate the proposed method on real collections of text documents and the experimental results show that our method achieves better performance comparing with others.