An Incremental Hierarchical Data Clustering Algorithm Based on Gravity Theory
One of the main challenges in the design of modern clustering algorithms is that, in many applications, new data sets are continuously added into an already huge database. As a result, it is impractical to carry out data clustering from scratch whenever there are new data instances added into the database. One way to tackle this challenge is to incorporate a clustering algorithm that operates incrementally. Another desirable feature of clustering algorithms is that a clustering dendrogram is generated. This feature is crucial for many applications in biological, social, and behavior studies, due to the need to construct taxonomies. This paper presents the GRIN algorithm, an incremental hierarchical clustering algorithm for numerical data sets based on gravity theory in physics. The GRIN algorithm delivers favorite clustering quality and generally features O(n) time complexity. One main factor that makes the GRIN algorithm be able to deliver favorite clustering quality is that the optimal parameters settings in the GRIN algorithm are not sensitive to the distribution of the data set. On the other hand, many modern clustering algorithms suffer unreliable or poor clustering quality when the data set contains highly skewed local distributions so that no optimal values can be found for some global parameters. This paper also reports the experiments conducted to study the characteristics of the GRIN algorithm.
Keyworddata clustering hierarchical clustering incremental clustering gravity theory
Unable to display preview. Download preview PDF.
- 1.M. Charikar, C. Chekuri, T. Feder and R. Motwani: Incremental Clustering and Dynamic Information Retrieval. In Proceedings of the 29th Annual ACM Symposium on Theory of Computing (STOC-97), 1997, pp. 626–634.Google Scholar
- 2.M. Ester, H.-P. Kriegel, J. Sander, M. Wimmer, and X. Xu. Incremental clustering for mining in a data warehousing environment. In Proceedings of 24th International Conference on Very Large Data Bases (VLDB-98), 1998, pp. 323–333.Google Scholar
- 4.D. Fisher, Improving inference through conceptual clustering, In Proceedings of 6th National Conference on Artificial Intelligence (AAAI-87), 1987, pp. 461–465.Google Scholar
- 6.J. Han, M. Kamber, Data Mining: Concepts and Techniques, San Francisco: Morgan Kaufmann Publishers, 2000.Google Scholar
- 7.R. V. Hogg and E. A. Tanis, Probability and statistical inference, New Jersey: Prentice Hall, 2001.Google Scholar
- 10.Yen-Jen Oyang, Chien-Yu Chen, and Tsui-Wei Yang, A Study on the Hierarchical Data Clustering Algorithm Based on Gravity Theory, In Proceedings of 5th European Conference on Principles and Practice of Knowledge Discovery in Databases (PKDD-01), 2001, pp. 350–361.Google Scholar
- 11.Yen-Jen Oyang, Chien-Yu Chen, Shien-Ching Hwang, and Cheng-Fang Lin, Characteristics of a Hierarchical Data Clustering Algorithm Based on Gravity Theory, Technical Report of NTUCSIE 02-01. (Available at http://mars.csie.ntu.edu.tw/~cychen/publications_on_dm.htm)
- 12.A. Ribert, A. Ennaji, and Y. Lecourtier, An incremental Hierarchical Clustering, In Proceedings of 1999 Vision Interface Conference, 1999, pp. 586–591.Google Scholar
- 13.M. Stonebraker, J. Frew, K. Gardels and J. Meredith, The Sequoia 2000 Storage Benchmark, In Proceedings of 1993 ACM-SIGMOD International Conference on Management of Data (SIGMOD-93), 1993, pp. 2–11.Google Scholar
- 14.I. H. Witten, Data mining: practical machine learning tools and techniques with Java implementations, San Francisco, Califonia: Morgan Kaufmann, 2000.Google Scholar
- 15.T. Zhang, R. Ramakrishnan, M. Livny, BIRCH: An Efficient Data Clustering Method for Very Large Databases, In Proceedings of the 1996 ACM-SIGMOD International Conference on Management of Data (SOGMOD-96), Jun. 1996, pp. 103–114.Google Scholar