Matrix Learning in Learning Vector Quantization

@inproceedings{Biehl2006MatrixLI,
  title={Matrix Learning in Learning Vector Quantization},
  author={Michael Biehl and Barbara Hammer and Petra Schneider},
  year={2006}
}
We propose a new matrix learning scheme to extend Generalize d Relevance Learning Vector Quantization (GRLVQ), an efficient prototype-ba sed classification algorithm. By introducing a full matrix of relevance factors in t he distance measure, correlations between different features and their importa nce for the classification scheme can be taken into account and automated, general metr ic adaptation takes place during training. In comparison to the weighted euclid ean metric used for GRLVQ, a… CONTINUE READING

References

Publications referenced by this paper.
Showing 1-10 of 18 references

and T

  • B. Hammer, M. Strickert
  • Villmann. On the general iz tion ability of GR…
  • 2005
Highly Influential
4 Excerpts

On the generaliz tion ability of GRLVQ networks

  • M. Strickert B. Hammer, T. Villmann
  • Neural Processing Letters
  • 2005

Obermayer , Soft Learning Vector Quantization

  • K.
  • 2003

Similar Papers

Loading similar papers…