Using Manifold Structure for Partially Labelled Classification

@inproceedings{ElkinUsingMS,
  title={Using Manifold Structure for Partially Labelled Classification},
  author={M Elkin and Partha Niyogi}
}
We consider the general problem of utilizing both labeled and un-labeled data to improve classification accuracy. Under t he assumption that the data lie on a submanifold in a high dimensional space, we develop an algorithmic framework to classify a partially labeled data set in a principled manner. The central idea of our approach is that classification functions are naturally defined only on t he sub-manifold in question rather than the total ambient space. Using the Laplace Beltrami operator… CONTINUE READING
Highly Influential
This paper has highly influenced 18 other papers. REVIEW HIGHLY INFLUENTIAL CITATIONS
Highly Cited
This paper has 218 citations. REVIEW CITATIONS

From This Paper

Figures, tables, and topics from this paper.

Citations

Publications citing this paper.
Showing 1-10 of 119 extracted citations

219 Citations

0102030'05'08'11'14'17
Citations per Year
Semantic Scholar estimates that this publication has 219 citations based on the available data.

See our FAQ for additional information.

References

Publications referenced by this paper.
Showing 1-9 of 9 references

Lap lacian Eigenmaps for Dim ensionality R edu ction and Data R epresentation

  • M Belkin, P Niyogi
  • Lap lacian Eigenmaps for Dim ensionality R edu…
  • 2002
Highly Influential
3 Excerpts

N onlin ear Dimensionality Reduction by Locally Linear Embedding

  • T Sam, Lawrence K Roweis, Saul
  • Science
  • 2000

Text Classifi cation from Labeled in Unlabeled Data

  • K Nigam, A K Mccallum, S Thrun, T Mitchell
  • Machine Learning
  • 2000
1 Excerpt

Spectra l Graph Theory

  • R K Fan, Chung
  • Regional Conference Series in Mathematics
  • 1997
1 Excerpt

The Laplacian on a Riemmannian Manifold

  • S Rosenberg
  • The Laplacian on a Riemmannian Manifold
  • 1997
2 Excerpts

Number of Labeled Points Figure 3: 20 Newsgroups data set. Error rates for different numbers of labeled and unlab eled points compared to best k-NN baseline

  • Number of Labeled Points Figure 3: 20 Newsgroups…
  • 1800

Similar Papers

Loading similar papers…