Using Domain-Specific Knowledge in SVMs Generalization Error Bounds

Eryarsoy, Enes and Koehler, Gary J. and Aytuğ, Haldun (2007) Using Domain-Specific Knowledge in SVMs Generalization Error Bounds. (Submitted)

This is the latest version of this item.

[img]PDF - Registered users only - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader


In this study we describe a methodology to exploit a specific type of domain knowledge in order to find tighter error bounds on the performance of classification via Support Vector Machines. The domain knowledge we consider is that the input space lies inside of a specified convex polytope. First, we con-sider prior knowledge about the domain by incorporating upper and lower bounds of attributes. We then consider a more general framework that allows us to encode prior knowledge in the form of linear con-straints formed by attributes. By using the ellipsoid method from optimization literature, we show that, this can be exploited to upper bound the radius of the hyper-sphere that contains the input space, and en-ables to tighten generalization error bounds. We provide a comparative numerical analysis and compare the effectiveness of our approach.

Item Type:Article
Uncontrolled Keywords:Prior knowledge; support vector machines; ellipsoid method; error bounds.
Subjects:Q Science > QA Mathematics > QA075 Electronic computers. Computer science
Q Science > QA Mathematics
ID Code:6596
Deposited By:Enes Eryarsoy
Deposited On:31 Oct 2007 09:25
Last Modified:15 Sep 2009 14:23

Available Versions of this Item

Repository Staff Only: item control page