Sciweavers

Free Online Productivity Tools i2Speak i2Symbol i2OCR iTex2Img iWeb2Print iWeb2Shot i2Type iPdf2Split iPdf2Merge i2Bopomofo i2Arabic i2Style i2Image i2PDF iLatex2Rtf Sci2ools

12

ICML
2005
IEEE

favoriteEmaildiscussreport

137views Machine Learning» more ICML 2005»

Large margin non-linear embedding

14 years 5 months ago

Large margin non-linear embedding

Download www.kyb.mpg.de

It is common in classification methods to first place data in a vector space and then learn decision boundaries. We propose reversing that process: for fixed decision boundaries, we "learn" the location of the data. This way we (i) do not need a metric (or even stronger structure) ? pairwise dissimilarities suffice; and additionally (ii) produce low-dimensional embeddings that can be analyzed visually. We achieve this by combining an entropybased embedding method with an entropybased version of semi-supervised logistic regression. We present results for clustering and semi-supervised classification.

Alexander Zien, Joaquin Quiñonero Candela

Real-time Traffic

Decision Boundaries | ICML 2005 | Machine Learning | Produce Low-dimensional Embeddings | Semi-supervised Logistic Regression |

claim paper

Related Content

» The Evolutionary Design and Synthesis of NonLinear Digital VLSI Systems

» Exploring Large Feature Spaces with Hierarchical Multiple Kernel Learning

» KLocal Hyperplane and Convex Distance Nearest Neighbor Algorithms

» Limitations of Learning via Embeddings in Euclidean HalfSpaces

» Learning NonMetric Partial Similarity Based on Maximal Margin Criterion

» Graph Embedding A General Framework for Dimensionality Reduction

» ClassMap Efficient Multiclass Recognition via Embeddings

» Graph Embedding and Extensions A General Framework for Dimensionality Reduction

» Better multiclass classification via a marginoptimized single binary problem

Post Info
More Details (n/a)

Added	17 Nov 2009
Updated	17 Nov 2009
Type	Conference
Year	2005
Where	ICML
Authors	Alexander Zien, Joaquin Quiñonero Candela

Comments (0)