Geometrical properties of Nu support vector machines with different norms. (English) Zbl 1075.68074

Summary: By employing the \(L_1\) or \(L_{\infty}\) norms in maximizing margins, Support Vector Machines (SVMs) result in a linear programming problem that requires a lower computational load compared to SVMs with the \(L_2\) norm. However, how the change of norm affects the generalization ability of SVMs has not been clarified so far except for numerical experiments. In this letter, the geometrical meaning of SVMs with the \(L_p\) norm is investigated, and the SVM solutions are shown to have rather little dependency on \(p\).


68T05 Learning and adaptive systems in artificial intelligence
Full Text: DOI


[1] DOI: 10.1016/S0925-2312(03)00380-1 · Zbl 02060369
[2] Cortes C., Machine Learning 20 pp 273– (1995)
[3] DOI: 10.1162/089976604774201659 · Zbl 1102.68561
[4] DOI: 10.1002/scj.10629 · Zbl 05440822
[5] Ikeda K., ICSC. NC pp 2– (2004)
[6] DOI: 10.1016/j.neunet.2004.11.008 · Zbl 1071.68538
[7] DOI: 10.1016/S0167-6377(98)00049-2 · Zbl 1028.90037
[8] DOI: 10.1016/S0167-8655(01)00071-X · Zbl 0990.68117
[9] DOI: 10.1162/089976600300015565
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.