Application of Self-Organizing Maps to Outlier Identification and Estimation of Missing Data

  • Mariusz Grabowski
Part of the Studies in Classification, Data Analysis, and Knowledge Organization book series (STUDIES CLASS)


Kohonen’s self-organizing maps (SOM) belong to the group of artificial neural network methods that are the most frequently applied to data analysis. The most common applications of SOM are multidimensional data visualization and huge data sets clustering. Some characteristics of SOM make this method interesting also in other aspects of data analysis. The following paper presents the possibility of SOM application to outliers identification and missing data estimation.


Neural Networks Self-organizing Maps Outliers Missing Data 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. Studies of Self-organizing Map (SOM) and Learning Vector Quantization (LVQ), (1997), Compiled in the Laboratory of Computer and Information Science, Helsinki University of Technology, Scholar
  2. Cherkassky V., Lari-Najafi H. (1991). Constrained Topological Mapping for Nonparametric Regression Analysis, Neural Networks, Vol. 4. pp. 27–40CrossRefGoogle Scholar
  3. Gałuszka B. (1994), Szacowanie brakujących danych w szeregach czasowych (wyniki badan), in: Przestrzenno-czasowe modelowanie i prognozowanie zjawisk gospodarczych, Aleksander Zelias, Krakow, pp. 134–141Google Scholar
  4. Gałuszka B., (1995), Szacowanie brakujących informacji w szeregach przekrojowych (wyniki badan), in.: Przestrzenno-czasowe modelowanie i prognozowanie zjawisk gospodarczych, Aleksander Zelias, Krakow, pp. 199–204Google Scholar
  5. Ghahramani Z., Jordan M.I. (1994). Learning From Incomplete Data, MIT, 500-1999/ Scholar
  6. Jajuga K. (1993). Statystyczna analiza wielowymiarowa, PWN, WarszawaGoogle Scholar
  7. Kohonen T. (1995). Self-Organizing Maps, Springer-Verlag BerlinGoogle Scholar
  8. Kohonen T., Hynninen J., Kangas J., Laaksonen J. (1995). SOM_PAK The Self- Organizing Map Program Package Version 3.1, Helsinki University of Technology, Scholar
  9. Kordos J. (1988). Jakość danych statystycznych, PWE, WarszawaGoogle Scholar
  10. Murtagh F. (1995). Interpreting Kohonen self-organizing map using contiguity- constrained clustering, Pattern Recognition Letters, 16, pp. 399–408Google Scholar
  11. Murtagh F. Hernändez-Pajeraz (1995a). The Kohonen self-organizing map method: an assessment, Journal of Classification, 12, pp. 165–190Google Scholar
  12. Murtagh F., Hernändez-Pajeraz (1995b). Clustering moderately-sized datasets using the Kohonen map approach, Statistics in Transition - Journal of the Polish Statistical Association, 2, pp. 151–162Google Scholar
  13. Oja E., Karhunen J., Wang L., Vigario R., (1995). Principal and independent components in neural networks - recent developments, Proc. VII Italian Workshop on Neural Nets WIRN’95, May 18 - 20, Vietri sul Mare, ItalyGoogle Scholar
  14. Sarle W. S. (1994). Neural Networks and Statistical Models, Proceedings of the Nineteenth Annual SAS Users Group International Conference, AprilGoogle Scholar
  15. Ultsh A. (1993). Self organized feature maps monitoring and knowledge aquisition of a chemical process, in: Procedings of the International Conference on Artificial Neural Networks (ICANN93), S. Gielen, B. Kappen, pp. 864–867, Springer-Verlag, London.Google Scholar

Copyright information

© Springer-Verlag Berlin · Heidelberg 1998

Authors and Affiliations

  • Mariusz Grabowski
    • 1
  1. 1.Department of Computer ScienceCracow University of EconomicsCracowPoland

Personalised recommendations