High-dimensional Data Analysis: From Optimal Metrics to Feature Selection Export

@book{citeulike:4502556, abstract = {High-dimensional data are everywhere: texts, sounds, spectra, images, etc. However, many data analysis tools (coming from statistics, artificial intelligence, etc.) were designed for low-dimensional data. Many of the assumptions behind data analysis tools are not transposable to high- dimensional data. For instance, the Euclidean distance concentrates in high- dimensional spaces; all distances seem identical! It furthermore does not distinguish between relenvant and irrelevant features. In Part One of the book, the phenomenon of the concentration of the distances is considered, and its consequences on data analysis tools are studied. Part Two focuses on the problem of feature selection in the case of a large number of initial features. Most of the concepts studied and presented in this thesis are illustrated on chemometric data, and more particularly on spectral data, with the objective of inferring a physical or chemical property of a material by analysis the spectrum of the light it reflects.}, author = {Fran\c{c}ois, Damien}, citeulike-article-id = {4502556}, citeulike-linkout-0 = {http://www.amazon.ca/exec/obidos/redirect?tag=citeulike09-20&path=ASIN/3836493098}, citeulike-linkout-1 = {http://www.amazon.de/exec/obidos/redirect?tag=citeulike01-21&path=ASIN/3836493098}, citeulike-linkout-2 = {http://www.amazon.fr/exec/obidos/redirect?tag=citeulike06-21&path=ASIN/3836493098}, citeulike-linkout-3 = {http://www.amazon.jp/exec/obidos/ASIN/3836493098}, citeulike-linkout-4 = {http://www.amazon.co.uk/exec/obidos/ASIN/3836493098/citeulike00-21}, citeulike-linkout-5 = {http://www.amazon.com/exec/obidos/redirect?tag=citeulike07-20&path=ASIN/3836493098}, citeulike-linkout-6 = {http://www.worldcat.org/isbn/3836493098}, citeulike-linkout-7 = {http://books.google.com/books?vid=ISBN3836493098}, citeulike-linkout-8 = {http://www.amazon.com/gp/search?keywords=3836493098&index=books&linkCode=qs}, citeulike-linkout-9 = {http://www.librarything.com/isbn/3836493098}, day = {06}, howpublished = {Paperback}, isbn = {3836493098}, keywords = {concentration-of-measure, high-dimensionality, kernel-machines, local-learning, locality, nearest-neighbors, overfitting}, month = {May}, posted-at = {2009-05-11 12:49:48}, priority = {2}, publisher = {VDM Verlag}, title = {High-dimensional Data Analysis: From Optimal Metrics to Feature Selection}, url = {http://www.amazon.com/exec/obidos/redirect?tag=citeulike07-20&path=ASIN/3836493098}, year = {2008} }

TY - BOOK ID - citeulike:4502556 L3 - citeulike-article-id:4502556 N2 - High-dimensional data are everywhere: texts, sounds, spectra, images, etc. However, many data analysis tools (coming from statistics, artificial intelligence, etc.) were designed for low-dimensional data. Many of the assumptions behind data analysis tools are not transposable to high- dimensional data. For instance, the Euclidean distance concentrates in high- dimensional spaces; all distances seem identical! It furthermore does not distinguish between relenvant and irrelevant features. In Part One of the book, the phenomenon of the concentration of the distances is considered, and its consequences on data analysis tools are studied. Part Two focuses on the problem of feature selection in the case of a large number of initial features. Most of the concepts studied and presented in this thesis are illustrated on chemometric data, and more particularly on spectral data, with the objective of inferring a physical or chemical property of a material by analysis the spectrum of the light it reflects. SN - 3836493098 TI - High-dimensional Data Analysis: From Optimal Metrics to Feature Selection PB - VDM Verlag KW - concentration-of-measure KW - high-dimensionality KW - kernel-machines KW - local-learning KW - locality KW - nearest-neighbors KW - overfitting AU - François, Damien PY - 2008/05/06/ UR - http://www.amazon.ca/exec/obidos/redirect?tag=citeulike09-20&path=ASIN/3836493098 ER -