Skip to main content
Log in

A weighted SOM for classifying data with instance-varying importance

  • Original Article
  • Published:
International Journal of Machine Learning and Cybernetics Aims and scope Submit manuscript


This paper presents a weighted self-organizing map (WSOM) that combines the advantages of the standard SOM paradigm with learning that accounts for instance-varying importance. While the learning of the classical batch SOM weights data by a neighborhood function, it is here augmented with a user-specified instance-specific importance weight for cost-sensitive classification. By focusing on instance-specific importance to the learning of a SOM, we take a perspective that goes beyond the common approach of incorporating a cost matrix into the objective function of a classifier. This paper provides evidence of the performance of the WSOM on standard benchmark and real-world data. We compare the WSOM with a classical SOM and a conventional statistical approach in two financial classification tasks: credit scoring and financial crisis prediction. The significance of instance-varying importance weights, and the performance of the WSOM, is confirmed by being superior in terms of cost-sensitive classifications.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save

Springer+ Basic
EUR 32.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or Ebook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Fig. 1
Fig. 2

Similar content being viewed by others


  1. Weighting would also be applicable for the sequential algorithm. The weight should, however, be applied to the learning rate α rather than to each data point in (4). Kohonen [18] further reminds that αW < 1 to guarantee stability, which implies that weighting would not be applicable during the first training cycles with large α. This is not, however, a concern with the batch algorithm.


  1. Barreto G (2007) Time series prediction with the self-organizing map: A review. In: Hitzler P, Hammer B (eds) Perspectives on neural-symbolic integration. Springerg, Berlin, pp 135–158

    Chapter  Google Scholar 

  2. Blake C, Merz C (1998) UCI repository of machine learning databases.

  3. Candelon B, Dumitrescu E, Hurlin C (2012) How to evaluate an early warning system? Towards a unified statistical framework for assessing financial crises forecasting methods. IMF Econ Rev 60(1):75–113

    Article  Google Scholar 

  4. Chappell G, Taylor J (1993) The temporal Kohonen map. Neural Netw 6:441–445

    Article  Google Scholar 

  5. Elkan C (2001) The foundations of cost-sensitive learning. In: Proceedings of the international joint conference on artificial intelligence (IJCAI 01), pp 973–978

  6. Fawcett F (2006) ROC graphs with instance-varying costs. Pattern Recogn Lett 27(8):882–891

    Article  MathSciNet  Google Scholar 

  7. Fawcett F (2008) PRIE: a system for generating rulelists to maximize ROC performance. Data Min Knowl Disc 17(2):207–224

    Article  MathSciNet  Google Scholar 

  8. Fawcett T, Foster J (1997) Provost: adaptive fraud detection. Data Min Knowl Disc 1(3):291–316

    Article  Google Scholar 

  9. Forte JC, Letrémy P, Cottrell M (2002) Advantages and drawbacks of the Batch Kohonen algorithm. In: Proceedings of the European symposium on artificial neural networks (ESANN 02), Springer, Berlin, pp 223–230

  10. Fuertes A-M, Kalotychou E (2006) Early warning systems for sovereign debt crises: the role of heterogeneity. Comput Stat Data Anal 51(2):1420–1441

    Article  MATH  MathSciNet  Google Scholar 

  11. Fuertes A-M, Kalotychou E (2007) Towards the optimal design of an early warning system for sovereign debt crises. Int J Forecast 23(1):85–100

    Article  Google Scholar 

  12. Hand DJ (2009) Mining the past to determine the future: problems and possibilities. Int J Forecast 25(3):441–451

    Article  Google Scholar 

  13. Hollmén J, Skubacz M (2000) Input dependent misclassification costs for cost-sensitive classifiers. In: Proceedings of the international conference on data mining

  14. Kangas J (1995) Sample weighting when training self-organizing maps for image compression. In: Proceedings of the 1995 IEEE workshop on neural networks for signal processing, pp 343–350

  15. Kaski S, Honkela T, Lagus K, Kohonen T (1998) WEBSOM—self-organizing maps of document collections. Neurocomputing 21:101–117

    Article  MATH  Google Scholar 

  16. Kim KY, Ra JB (1993) Edge preserving vector quantization using self-organizing map based on adaptive learning. In: Proceedings of the international joint conference on neural networks (IJCNN 93), vol 11. IEEE Press, pp 1219–1222

  17. Kohonen T (1991) The Hypermap architecture. In: Kohonen T, Mäkisara K, Simula O, Kangas J (eds) Artificial neural networks, vol II. Elsevier, Amsterdam, pp 1357–1360

    Google Scholar 

  18. Kohonen T (1993) Things you haven’t heard about the Self-Organizing Map. In: Proceedings of the international conference on neural networks (ICNN 93), pp 1147–1156

  19. Kohonen T (2001) Self-organizing maps, 3rd edn. Springer, Berlin

    Book  MATH  Google Scholar 

  20. Kumar M, Moorthy U, Perraudin W (2003) Predicting emerging market currency crashes. J Empir Finance 10(4):427–454

    Article  Google Scholar 

  21. Lo Duca M, Peltonen T (2013) Assessing systemic risks and predicting systemic events. J Banking Finance 37(7):2183–2195

    Google Scholar 

  22. Lomax S, Vadera S (2013) A survey of cost-sensitive decision tree induction algorithms. ACM Comput Surv 45(2):16:1–16:35

    Google Scholar 

  23. Reinhart CM, Rogoff KS (2008) Is the 2007 US sub-prime financial crisis so different? An international historical comparison. Am Econ Rev 98(2):339–344

    Article  Google Scholar 

  24. Reinhart CM, Rogoff KS (2009) The aftermath of financial crises. Am Econ Rev 99(2):466–472

    Article  Google Scholar 

  25. Sarlin P (2012a) Data and dimension reduction for visual financial performance analysis. TUCS Technical Report 1049, May 2012

  26. Sarlin P (2012b) Visual tracking of the millennium development goals with a fuzzified self-organizing neural network. Int J Mach Learn Cybern 3(3):233–245

    Article  Google Scholar 

  27. Sarlin P (2013a) On policymakers’ loss functions and the evaluation of early warning systems. Econ Let 119(1):1–7

    Google Scholar 

  28. Sarlin P (2013b) Self-organizing time map: an abstraction of temporal multivariate patterns. Neurocomputing 99(1):496–508

    Article  Google Scholar 

  29. Sarlin P, Peltonen TA (2013) Mapping the state of financial stability. J Int Financial Mark Inst Money. doi:10.1016/j.intfin.2013.05.002

  30. Vesanto J, Himberg J, Alhoniemi E, Parhankangas J (1999) Self-organizing map in Matlab: the SOM Toolbox. In: Proceedings of the Matlab DSP conference, pp 35–40

  31. Yao Z, Sarlin P, Eklund T, Back B (2012) Combining visual customer segmentation and response modeling. IN: Proceedings of the European conference on information systems (ECIS 12), June 2012

  32. Zadrozny B, Elkan C (2001) Learning and making decisions when costs and probabilities are both unknown. In: Proceedings of the ACM SIGKDD international conference on knowledge discovery and data mining (KDD 01), pp 204–213

Download references

Author information

Authors and Affiliations


Corresponding author

Correspondence to Peter Sarlin.

Rights and permissions

Reprints and permissions

About this article

Cite this article

Sarlin, P. A weighted SOM for classifying data with instance-varying importance. Int. J. Mach. Learn. & Cyber. 5, 101–110 (2014).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: