×

zbMATH — the first resource for mathematics

Lo Gerfo, L.; Rosasco, L.; Odone, F.; de Vito, E.; Verri, A.
Spectral algorithms for supervised learning. (English) Zbl 1147.68643
Neural Comput. 20, No. 7, 1873-1897 (2008).
Summary: We discuss how a large class of regularization methods, collectively known as spectral regularization and originally designed for solving ill-posed inverse problems, gives rise to regularized learning algorithms. All of these algorithms are consistent kernel methods that can be easily implemented. The intuition behind their derivation is that the same principle allowing for the numerical stabilization of a matrix inversion problem is crucial to avoid overfitting. The various methods have a common derivation but different computational and theoretical properties. We describe examples of such algorithms, analyze their classification performance on several data sets and discuss their applicability to real-world problems.

Cited in 25 Documents
MSC:
68T05 Learning and adaptive systems in artificial intelligence
Keywords:
regularization methods; spectral regularization
PDF BibTeX XML Cite
\textit{L. Lo Gerfo} et al., Neural Comput. 20, No. 7, 1873--1897 (2008; Zbl 1147.68643)
Full Text: DOI
References:
[1] DOI: 10.2307/1990404 · Zbl 0037.20701 · doi:10.2307/1990404
[2] DOI: 10.1198/016214505000000907 · Zbl 1118.62330 · doi:10.1198/016214505000000907
[3] DOI: 10.1016/j.jco.2006.07.001 · Zbl 1109.68088 · doi:10.1016/j.jco.2006.07.001
[4] DOI: 10.1162/153244302760200704 · Zbl 1007.68083 · doi:10.1162/153244302760200704
[5] DOI: 10.1198/016214503000125 · Zbl 1041.62029 · doi:10.1198/016214503000125
[6] DOI: 10.1007/s10208-006-0196-8 · Zbl 1129.68058 · doi:10.1007/s10208-006-0196-8
[7] DOI: 10.1007/s10208-004-0134-1 · Zbl 1083.68106 · doi:10.1007/s10208-004-0134-1
[8] DOI: 10.1142/S0219530506000711 · Zbl 1088.65056 · doi:10.1142/S0219530506000711
[9] De Vito E., Journal of Machine Learning Research 6 pp 883– (2005)
[10] DOI: 10.1023/A:1018946025316 · Zbl 0939.68098 · doi:10.1023/A:1018946025316
[11] DOI: 10.1162/neco.1995.7.2.219 · doi:10.1162/neco.1995.7.2.219
[12] Hastie S. T., JMLR 5 pp 1391– (2004)
[13] Micchelli C. A., JMLR 7 pp 2651– (2006)
[14] DOI: 10.1038/nature02341 · doi:10.1038/nature02341
[15] DOI: 10.1162/089976698300017575 · doi:10.1162/089976698300017575
[16] DOI: 10.1142/S0219530505000650 · Zbl 1101.68621 · doi:10.1142/S0219530505000650
[17] DOI: 10.1090/S0273-0979-04-01025-0 · Zbl 1107.94007 · doi:10.1090/S0273-0979-04-01025-0
[18] DOI: 10.1016/j.acha.2005.03.001 · Zbl 1107.94008 · doi:10.1016/j.acha.2005.03.001
[19] DOI: 10.1007/s00365-006-0659-y · Zbl 1127.68088 · doi:10.1007/s00365-006-0659-y
[20] DOI: 10.1007/s10208-004-0155-9 · Zbl 1100.68100 · doi:10.1007/s10208-004-0155-9
[21] DOI: 10.1007/s00365-006-0663-2 · Zbl 1125.62035 · doi:10.1007/s00365-006-0663-2
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.