Semantic mapping and K-means applied to hybrid SOM-based document organization system construction

ABSTRACT

In this paper, we present and evaluate a hybrid document organization system based on Self-Organizing Maps. The proposed system uses Semantic Mapping to dimensionality reduction and K-means to volume reduction of document vectors of a medium text collection. The vectors obtained after dimensionality and volume reduction steps are used to train the document maps with the SOM algorithm, thus the training time is reduced without compromising the quality of the generated map. We compare experimentally the hybrid system with the correspondent SOM system in organization of documents of Reuters-21758 v1.0 collection. The performances of the systems were measured in terms of classification error in text categorization and training time. The experimental results show that the proposed system generates pretty good document maps with smallest training time.

REFERENCES

• 1: T. Kohonen, S. Kaski, K. Lagus, J. Salojärvi, J. Honkela, V. Paatero and A. Saarela, Self Organization of a Massive Document Collection, IEEE Transaction on Neural Networks, v. 11, n. 3, 2000, pp. 574--585.
• 2: S. Kaski, Dimensionality Reduction by Random Mapping: Fast Similarity Computation for Clustering, In Proc. IJCNN'98, v. 1, 1998, 413--418.
• 3: Arnulfo P. Azcarraga , Teddy N. Yap, Jr., SOM-Based Methodology for Building Large Text Archives, Proceedings of the 7th International Conference on Database Systems for Advanced Applications, p.66-73, April 18-20, 2001
• 4: R. F. Corrêa, T. B. Ludermir, Dimensionality Reduction by Semantic Mapping, In Proc. VIII Brazilian Neural Networks Symposium. Los Alamitos: IEEE Computer Society, 2004.
• 5: R. F. Corrêa, T. B. Ludermir, Improving Self-Organization of Document Collections by Semantic Mapping, Neurocomputing (Amsterdam), v.70, 2006, 62--69.
• 6: Renato Fernandes Correa , Teresa Bernarda Ludermir, A Hybrid SOM-Based Document Organization System, Proceedings of the Ninth Brazilian Symposium on Neural Networks, p.16, October 23-27, 2006 [doi>10.1109/SBRN.2006.3]
• 7: Fabrizio Sebastiani, Machine learning in automated text categorization, ACM Computing Surveys (CSUR), v.34 n.1, p.1-47, March 2002 [doi>10.1145/505282.505283]
• 8: R. F. Corrêa, T. B. Ludermir, Dimensionality Reduction of very large document collections by Semantic Mapping. In Proc. 6th Workshop on Self-Organizing Maps WSOM 2007, Bielefeld, Germany, in press, 2007.
• 9: A. K. Jain , M. N. Murty , P. J. Flynn, Data clustering: a review, ACM Computing Surveys (CSUR), v.31 n.3, p.264-323, Sept. 1999 [doi>10.1145/331499.331504]
• 10: A. Strehl, J. Ghosh, R. Mooney, Impact of similarity measures on web-page clustering, In Proc. AAAI Workshop on AI for Web Search, 2000, 58--64.
• 11: R. R. Wilcox, Fundamentals of Modern Statistical Methods: Substantially Improving Power and Accuracy, Springer-Verlag, New York, 2001.
• 12: D. D. Lewis, Reuters-21578 Text Categorization Test Colletion, AT&T Labs Research, 1997. Available: http://www.research.att.com/~lewis.
• 13: Franca Debole , Fabrizio Sebastiani, An analysis of the relative hardness of Reuters-21578 subsets: Research Articles, Journal of the American Society for Information Science and Technology, v.56 n.6, p.584-596, April 2005 [doi>10.1002/asi.v56:6]

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.3 INFORMATION STORAGE AND RETRIEVAL
      H.3.3 Information Search and Retrieval
          Subjects: Clustering

General Terms:
Algorithms, Design, Experimentation, Performance

Keywords:
K-means, document classification and categorization, hybrid system, self-organizing maps, semantic mapping