• TUCS Publication Series
• How to Publish in TUCS Series
• Publication Search
• Publication Graph
• Publication Input
• Publication Input Guide
• JuFo Browser

A Weighted SOM for Classifying Data with Instance-Varying Importance

Peter Sarlin, A Weighted SOM for Classifying Data with Instance-Varying Importance. In: Jilles Vreeken, Charles Ling (Eds.), Proceedings of the IEEE 12th International Conference on Data Mining Workshops (ICDMW), 187–193 , IEEE, 2013.

Abstract:

This paper presents a Weighted Self-Organizing Map (WSOM). The WSOM 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, we augment it 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. When setting the weight to be the importance of an instance for forming clusters, the WSOM may also be seen as an alternative for cost-sensitive unsupervised clustering. We compare the WSOM with a classical SOM and logit analysis in financial crisis prediction. The performance of the WSOM in the financial setting is confirmed by superior cost-sensitive classification performance.

BibTeX entry:

@INPROCEEDINGS{inpSarlin_Peter13a,
  title = {A Weighted SOM for Classifying Data with Instance-Varying Importance},
  booktitle = {Proceedings of the IEEE 12th International Conference on Data Mining Workshops (ICDMW)},
  author = {Sarlin, Peter},
  editor = {Vreeken, Jilles and Ling, Charles},
  publisher = {IEEE},
  pages = {187–193 },
  year = {2013},
}

Belongs to TUCS Research Unit(s): Data Mining and Knowledge Management Laboratory

Edit publication