POPOVICI, Vlad, S BENGIO a JP THIRAN. Kernel matching pursuit for large datasets. PATTERN RECOGNITION. OXFORD: PERGAMON-ELSEVIER SCIENCE LTD, 2005, roč. 38, č. 12, s. 2385-2390. ISSN 0031-3203. Dostupné z: https://dx.doi.org/10.1016/j.patcog.2005.01.021. |
Další formáty:
BibTeX
LaTeX
RIS
@article{1090320, author = {Popovici, Vlad and Bengio, S and Thiran, JP}, article_location = {OXFORD}, article_number = {12}, doi = {http://dx.doi.org/10.1016/j.patcog.2005.01.021}, keywords = {kernel matching pursuit; greedy algorithm; sparse classifier}, language = {eng}, issn = {0031-3203}, journal = {PATTERN RECOGNITION}, title = {Kernel matching pursuit for large datasets}, volume = {38}, year = {2005} }
TY - JOUR ID - 1090320 AU - Popovici, Vlad - Bengio, S - Thiran, JP PY - 2005 TI - Kernel matching pursuit for large datasets JF - PATTERN RECOGNITION VL - 38 IS - 12 SP - 2385-2390 EP - 2385-2390 PB - PERGAMON-ELSEVIER SCIENCE LTD SN - 00313203 KW - kernel matching pursuit KW - greedy algorithm KW - sparse classifier N2 - Kernel matching pursuit is a greedy algorithm for building an approximation of a discriminant function as a linear combination of some basis functions selected from a kernel-induced dictionary. Here we propose a modification of the kernel matching pursuit algorithm that aims at making the method practical for large datasets. Starting from an approximating algorithm, the weak greedy algorithm, we introduce a stochastic method for reducing the search space at each iteration. Then we study the implications of using an approximate algorithm and we show how one can control the trade-off between the accuracy and the need for resources. Finally, we present some experiments performed on a large dataset that support our approach and illustrate its applicability. (c) 2005 Pattern Recognition Society. Published by Elsevier Ltd. All rights reserved. ER -
POPOVICI, Vlad, S BENGIO a JP THIRAN. Kernel matching pursuit for large datasets. \textit{PATTERN RECOGNITION}. OXFORD: PERGAMON-ELSEVIER SCIENCE LTD, 2005, roč.~38, č.~12, s.~2385-2390. ISSN~0031-3203. Dostupné z: https://dx.doi.org/10.1016/j.patcog.2005.01.021.
|