Soft Rank Neighbor Embeddings

Apr, 2013

Correlation-based multidimensional scaling is proposed for reconstructing pairwise dissimilarity or score relationships in a Euclidean space. Pearson correlation between pairs of objects in source and target space can be directly maximized by gradient methods, while gradient optimization of Spearman rank correlation profits from a numerically soft formulation introduced in this work. Scale and shift invariance properties of correlation help circumventing typical distance concentration problems

Link

@inproceedings{Strickert_ESANN_2013,
author = {Marc Strickert and Kerstin Bunte},
title = {Soft Rank Neighbor Embeddings},
booktitle = {Proc. of the  21th "European Symposium on Artificial Neural Networks (ESANN)},
address = {Bruges, Belgium},
year = {2013},
pages = {77--82},
publisher = {D-facto Publications},
owner = {strickert},
timestamp = {2011.06.23},
month = {"Apr."},
editor = {M. Verleysen},
isbn = {978-2-87419-081-0},
url = {http://www.elen.ucl.ac.be/Proceedings/esann/esannpdf/es2013-46.pdf},
abstract = {Correlation-based multidimensional scaling is proposed for reconstructing pairwise dissimilarity or score relationships in a Euclidean space.  Pearson correlation between pairs of objects in source and target space can be directly maximized by gradient methods,  while gradient optimization of Spearman rank correlation profits from a numerically soft formulation introduced in this work.  Scale and shift invariance properties of correlation help circumventing typical distance concentration problems},
}