Margin Analysis of the LVQ Algorithm

@inproceedings{Crammer2002MarginAO,
  title={Margin Analysis of the LVQ Algorithm},
  author={Koby Crammer and Ran Gilad-Bachrach and Amir Navot and Naftali Tishby},
  booktitle={NIPS},
  year={2002}
}
One of the earliest and most powerful machine learning methods is the Learning Vector Quantization (LVQ) algorithm, introduced by Kohonen about 20 years ago. Still, despite its popularity, the theoretical justification of this model is quite limited. In this paper we fill the gap and provide margin based analysis for this model. We present a rigorous bound on the generalization error which is independent of the dimension of the data. Furthermore we show that LVQ is a family of maximal margin… CONTINUE READING
Highly Influential
This paper has highly influenced 24 other papers. REVIEW HIGHLY INFLUENTIAL CITATIONS
Highly Cited
This paper has 282 citations. REVIEW CITATIONS

Citations

Publications citing this paper.
Showing 1-10 of 132 extracted citations

A novel feature selection algorithm based on LVQ hypothesis margin

Neural Computing and Applications • 2013
View 5 Excerpts
Highly Influenced

Neural Information Processing

Lecture Notes in Computer Science • 2012
View 20 Excerpts
Highly Influenced

Class Conditional Nearest Neighbor for Large Margin Instance Selection

IEEE Transactions on Pattern Analysis and Machine Intelligence • 2010
View 9 Excerpts
Highly Influenced

Feature selection based on loss-margin of nearest neighbor classification

Pattern Recognition • 2009
View 4 Excerpts
Highly Influenced

Relevance-Based Feature Extraction for Hyperspectral Images

IEEE Transactions on Neural Networks • 2008
View 4 Excerpts
Highly Influenced

Prototype based machine learning for clinical proteomics

Ausgezeichnete Informatikdissertationen • 2006
View 8 Excerpts
Highly Influenced

283 Citations

0102030'03'06'10'14'18
Citations per Year
Semantic Scholar estimates that this publication has 283 citations based on the available data.

See our FAQ for additional information.

References

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

The Nature of Statistical Learning Theory

Statistics for Engineering and Information Science • 2000
View 7 Excerpts
Highly Influenced

Self-Organizing Maps

View 4 Excerpts
Highly Influenced

Lvq is a maximum margin algorithm

L. Buckingham, S. Geva
Pacific Knowledge Acquisition Workshop PKAW’2000 • 2000
View 2 Excerpts

Machine learning for information retrieval: Advanced techniques

Y. Singer, D. D. Lewis
presented at ACM SIGIR 2000 • 2000
View 2 Excerpts

Support Vector MAchines

N. Cristianini, J. Shawe-Taylor
Cambridge University Press • 2000

Similar Papers

Loading similar papers…